bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes
1 CHD8 Regulates Cellular Homeostasis and Neuronal Function Genes 2 Across Multiple Models of CHD8 Haploinsufficiency 3 4 A. Ayanna Wade1,2, Kenneth Lim1,2, Rinaldo Catta-Preta1,2, Alex S. Nord1,2,* 5 6 1Department of Psychiatry and Behavioral Sciences, University of California, Davis, Davis, CA, 7 USA. 2Department of Neurobiology, Physiology and Behavior, University of California, Davis, 8 Davis, CA, USA. 9 10 *Correspondence: 11 Dr. Alex Nord 12 [email protected] 13 14 Word count: 5,132 15 Figure count: 15
1 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes
16 ABSTRACT 17 18 The packaging of DNA into chromatin determines the transcriptional potential of cells and is 19 central to eukaryotic gene regulation. Recent sequencing of patient mutations has linked de novo 20 loss-of-function mutations to chromatin remodeling factors with specific, causal roles in 21 neurodevelopmental disorders. Characterizing cellular and molecular phenotypes arising from 22 haploinsufficiency of chromatin remodeling factors could reveal convergent mechanisms of 23 pathology. Chromodomain helicase DNA binding protein 8 (CHD8) encodes a chromatin 24 remodeling factor gene and has among the highest de novo loss-of-function mutations rates in 25 patients with autism spectrum disorder (ASD). Mutations to CHD8 are expected to drive 26 neurodevelopmental pathology through global disruptions to gene expression and chromatin 27 state, however, mechanisms associated with CHD8 function have yet to be fully elucidated. We 28 analyzed published transcriptomic and epigenomic data across CHD8 in vitro and in vivo 29 knockdown and knockout models to identify convergent mechanisms of gene regulation by 30 CHD8. We found reproducible high-affinity interactions of CHD8 near promoters of genes 31 necessary for basic cell functions and gene regulation, especially chromatin organization and 32 RNA processing genes. Overlap between CHD8 interaction and differential expression suggests 33 that reduced dosage of CHD8 directly relates to decreased expression of these genes. In addition, 34 genes important for neuronal development and function showed consistent dysregulation, though 35 there was a reduced rate and decreased affinity for CHD8 interactions near these genes. This 36 meta-analysis verifies CHD8 as a critical regulator of gene expression and reveals a consistent 37 set of high affinity CHD8 interaction targets observed across human and mouse in vivo and in 38 vitro studies. Our findings highlight novel core functions of CHD8 and indicate direct and 39 downstream gene regulatory impacts that are likely to be associated with neuropathology 40 underlying CHD8-associated neurodevelopmental disorder. 41 42 Keywords: Autism spectrum disorder; CHD8; Chromatin remodeling; Functional genomics; 43 Neurodevelopment
2 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes
44 INTRODUCTION 45 46 Recent genetic studies suggest that single copy loss-of-function mutations to chromatin 47 remodeling genes significantly contribute to autism spectrum disorder (ASD) neurobiology, 48 presumably through disruptions to transcriptional regulation in the developing and mature brain 49 (De Rubeis et al. 2014, Iossifov et al. 2014, Parikshak et al. 2013, O’Roak et al. 2012a, O’Roak 50 et al. 2012b, Sanders et al. 2015, Vissers et al. 2016). The gene encoding chromodomain helicase 51 DNA binding protein 8 (CHD8) has one of the highest observed mutation rates in sporadic ASD 52 (O’Roak et al. 2012a, Barnard et al. 2015, Krumm et al. 2014), and mutations to CHD8 have also 53 been identified in cases from schizophrenia and intellectual disability cohorts (Tatton-Brown et 54 al. 2017, McCarthy et al. 2014). Patients that carry CHD8 mutations typically meet stringent 55 qualifications for ASD diagnosis, and frequently present with comorbid features including 56 macrocephaly, cognitive impairment, distinct craniofacial morphology, and gastrointestinal 57 disturbances (Bernier et al. 2014). Knockdown or haploinsufficiency of Chd8 in animal models 58 has recapitulated specific neuroanatomical, gastrointestinal, cognitive, and behavioral 59 phenotypes observed in patients (Sugathan et al. 2014, Gompers et al. 2017, Katayama et al. 60 2016, Platt et al. 2017), though reported phenotypes vary across models. Considering the 61 relevance of CHD8 mutations in neurodevelopmental disorders and the positive findings of 62 relevant phenotypes in model systems, characterizing the convergent patterns of CHD8 genomic 63 interactions and transcriptional outcomes caused by CHD8 haploinsufficiency across studies 64 could significantly advance understanding of core pathophysiology in patients carrying CHD8 65 mutations and, potentially, reveal generalized chromatin-associated cellular mechanisms 66 underlying neurodevelopmental disorders. 67 68 CHD8 belongs to a family of ATP-dependent chromatin remodelers (Hall and Georgel 69 2007, Marfella and Imbalzano 2007, Hargreaves and Crabtree 2011). CHD family proteins are 70 distinguished by tandem chromodomains predicted to enable these proteins to bind histones 71 (Flanagan et al. 2005). As some CHD proteins demonstrate chromatin remodeling activity (Hall 72 and Georgel 2007, McKnight et al. 2011, Tong et al. 1998), CHD8 has been speculated to drive 73 ASD-associated changes in neurodevelopmental gene expression by targeting and remodeling 74 chromatin at specific promoters and enhancers (Sugathan et al. 2014, Cotney et al. 2015, 75 Ceballos-Chavez et al. 2015). This is supported by evidence that CHD8 can reposition 76 nucleosomes in vitro and in mammalian cell culture (Thompson et al. 2008), and that loss of 77 Chd8 in in vitro and in vivo models dysregulates ASD-associated and Chd8-target gene 78 expression (Sugathan et al. 2014, Gompers et al. 2017, Katayama et al. 2016, Cotney et al. 79 2015). Several mechanisms have been suggested to underlie CHD8 binding specificity, including 80 targeting through histone modifications associated with open chromatin (Sugathan et al. 2014, 81 Cotney et al. 2015, Yuan et al. 2007, Rodriguez-Paredes et al. 2009) and recruitment through 82 protein-protein interactions (Thompson et al. 2008, Yuan et al. 2007, Rodriguez-Paredes et al. 83 2009, Ishihara et al. 2006, Nishiyama et al. 2009, Shen et al. 2015, Fang et al. 2016). However, 84 the mechanisms by which CHD8 directly regulates target gene expression, whether CHD8 85 targets cell- and stage-specific genes in the developing brain, and which patterns of 86 transcriptional dysregulation are due to direct effects versus downstream or secondary changes to 87 CHD8 regulation remain unresolved. 88
3 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes
89 There are a growing number of studies that have explored the role of CHD8 in 90 neurodevelopment, providing the opportunity to test for core features of CHD8 genomic 91 interactions and transcriptomic dysregulation associated with CHD8 haploinsufficiency. 92 Published studies have encompassed both in vitro and in vivo systems with shRNA knockdown 93 or genetic mutation of CHD8. Despite the variety of models there appear to be general patterns 94 of neurodevelopmental disruption caused by reduced CHD8 expression, characterized by 95 impacts to cellular proliferation, neuronal differentiation, and synaptic function. However, 96 discrepancies between cellular and behavioral findings make it difficult to reconcile core features 97 of cellular pathology. While published models of CHD8 haploinsufficiency vary considerably, 98 nearly all such studies have leveraged genomic approaches to determine the impact of CHD8 99 haploinsufficiency on gene expression. Many have also examined CHD8 interaction targets 100 genome-wide. The methods used for these experiments, RNA sequencing (RNA-seq) and 101 chromatin immunoprecipitation followed by sequencing (ChIP-seq) can generate comparable, 102 unbiased, and quantitative data enabling direct comparisons of results across models and studies. 103 We hypothesized that meta-analysis of these datasets using the same computational methods may 104 capture the consistent patterns of transcriptional pathology associated with CHD8 105 haploinsufficiency and reveal constitutive and model-specific genomic interaction patterns of 106 CHD8. 107 108 Here, we re-analyzed published RNA- and ChIP-seq data from CHD8 in vitro and in vivo 109 mouse and human models, and built an online user interface to enable customizable data analysis 110 and visualization across transcriptomic studies of CHD8 haploinsufficiency. Across studies, we 111 found a reproducible set of high-affinity CHD8 interaction target genes important for cellular 112 homeostasis, with many datasets additionally exhibiting downregulated gene expression of these 113 targets. We also found a secondary signature of transcriptional dysregulation of genes important 114 for neuronal development and function consistent with CHD8 haploinsufficiency. The findings 115 of this meta-analysis indicate evolutionarily-conserved functions of CHD8, with reductions in 116 CHD8 expression directly and indirectly altering transcription of genes critical for cellular 117 homeostasis and neurodevelopment in mouse and human models. 118 119 MATERIALS AND METHODS 120 121 CHD8 genomic datasets 122 123 Next generation sequencing datasets generated from CHD8 studies were identified through 124 a literature search of publications featuring the keyword “CHD8” in the PubMed and Gene 125 Expression Omnibus (GEO) databases. Raw data from publications that featured RNA-seq or 126 ChIP-seq analysis were downloaded from GEO with the exception of three publications that hosted 127 raw data on DDBJ (Katayama et al. 2016) and SRA (Platt et al. 2017, Wilkinson et al. 2015). A 128 total of twelve publications corresponding to 289 sequencing libraries were included in the 129 analysis. Libraries from Cotney et al. (2015) generated from fetal brain and libraries from Han et 130 al. (2017) designed for analysis of alternative splicing were not included in the analysis. All data 131 included were stated to be compliance with respective animal care and use committees at time of 132 original publication. 133 134
4 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes
135 RNA-seq analysis 136 137 RNA-seq computational analysis was performed following an established pipeline using 138 standard software, as described previously (Gompers et al. 2017). Briefly, unaligned sequencing 139 reads were assessed for general quality using FastQC (Version 0.11.2) and aligned to the mouse 140 (mm9) or human (GRCh37) reference genome using STAR (Version 2.5.2b, Dobin et al. 2013). 141 Aligned reads mapping to genes according to the mm9 genes.gtf or to gencode.v19.annotation.gtf 142 were counted at the gene level using subreads featureCounts (Version 1.5.0-p1, Liao et al. 2014). 143 Overall data quality, including testing for GC-bias, gene body coverage bias, and proportion of 144 reads in exons was further assessed using RSeQC (Version 2.6.4, Wang et al. 2012). Raw gene 145 count data and sample information as reported in the respective repositories were used for 146 differential expression analysis using edgeR (Version 3.4.4, Robinson et al. 2010). Genes with at 147 least 1 count per million were included in a general linearized model using a sequencing-run factor- 148 based covariate with genotype or knockdown as the variables for testing. For some datasets 149 additional covariates were included if described in the original publication. Where possible, overall 150 patterns of differentially expressed genes were compared to the original publication to ensure 151 consistency in results. Normalized expression levels were generated using the edgeR rpkm 152 function. Normalized log2(RPKM) values were used for plotting summary heatmaps and for 153 expression data of individual genes. Variation in sequencing depth and intra-study sample 154 variability partially account for differences in sensitivity and power across studies and likely drive 155 some of the differences observed across studies, including the total number of differentially 156 expressed genes. To capture an inclusive set of differentially expressed genes (DEGs), DEGs were 157 defined by uncorrected p-values < 0.05. DEG sets were used for gene set enrichment analysis for 158 Gene Ontology terms. 159 160 ChIP-seq analysis 161 162 ChIP-seq analysis was also performed using an established pipeline and standard methods, 163 as reported before (Gompers et al. 2017). Briefly, unaligned sequencing reads were assessed for 164 general quality using FastQC and mapped to the mouse (mm9) or human (hg19) genome using 165 BWA (Version 0.7.13, Li and Durbin 2009). Significant peaks with a p-value of < 0.0001 were 166 identified using MACS2 (Version 2.1.0, Feng et al. 2011) with model-based peak identification 167 and local significance testing disabled. Test datasets were analyzed comparing each individual 168 ChIP-seq experiment to matched input or IgG controls. Input and IgG libraries were analyzed 169 using the same approach to test for technical artifacts that could confound ChIP-seq results 170 generally following a previously reported quality control strategy (Marinov et al. 2014). Enriched 171 regions from IP and control datasets were annotated to genomic features using custom R scripts 172 and the combined UCSC and RefSeq transcript sets for the mouse or human genome build. CHD8 173 target genes were assigned by peak annotation to transcript start site (TSS) or to the nearest TSS 174 for distal peaks. HOMER was used to perform de novo motif discovery with default parameters 175 (Version 4.7, Heinz et al. 2010). Where possible, we verified that results from ChIP-seq reanalysis 176 were consistent with original publication. 177 178 179 180
5 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes
181 Gene ontology enrichment 182 183 The goseq R package (Version 1.30.0, Young et al. 2010) was used to test for enrichment 184 of gene ontology terms while correcting for gene length. Analysis included GO Biological Process, 185 Molecular Function, and Cellular Component annotations and required a minimal node size, or 186 number of genes annotated to GO terms, of 20. The internal ‘weight01’ testing framework and 187 Fishers test was used to account for multiple testing comparisons. Down- and upregulated genes 188 were examined separately for RNA-seq GO analysis using a goseq FDR < 0.05 cutoff. ChIP-seq 189 gene sets for the GO analysis were analyzed for experimental and control libraries separately also 190 with a goseq FDR < 0.05 cutoff. Test gene sets for DEGs and CHD8 interaction targets were 191 compared against a background set of expressed genes based on the minimum read-count cutoffs 192 for each dataset for DEGs or a background set of all conserved mouse-human genes identified 193 across RNA-seq datasets for CHD8 target genes. Heatmaps showing positive 194 log2(expected/observed) values were plotted for GO terms for data visualization. 195 196 Code and data availability and additional analysis visualization 197 198 Data that support the findings of this study are available from the corresponding author 199 upon request. Accession numbers in parentheses and DOIs for all published gene sets used in 200 enrichment analysis: 201 202 Ceballos-Chavez et al. (GSE62428): https://dx.doi.org/10.1371/journal.pgen.1005174; 203 Cotney et al. (GSE57369): https://dx.doi.org/10.1038/ncomms7404; 204 de Dieuleveult et al. (GSE64825): https://dx.doi.org/10.1038/nature16505; 205 Durak et al. (GSE72442): https://dx.doi.org/10.1038/nn.4400; 206 Gompers et al. (GSE99331): https://dx.doi.org/10.1038/nn.4592; 207 Katayama et al. (DRA003116): https://dx.doi.org/10.1038/nature19357; 208 Platt et al. (PRJNA379430): https://dx.doi.org/10.1016/j.celrep.2017.03.052; 209 Shen et al. (GSE71183, GSE71185): https://dx.doi.org/10.1016/j.molcel.2015.10.033; 210 Sugathan et al. (GSE61492): https://dx.doi.org/10.1073/pnas.1405266111; 211 Wang et al. 2015 (GSE71594): https://dx.doi.org/10.1186/s13229-015-0048-6; 212 Wang et al. 2017 (GSE85417): https://dx.doi.org/10.1186/s13229-017-0124-1; 213 Wilkinson et al. (PRJNA305612): https://dx.doi.org/10.1038/tp.2015.62; 214 215 Expanded results of the meta-analysis reported here are available from the interactive web 216 server at https://nordlab.shinyapps.io/rna_browser/. ChIP-seq datasets are available as UCSC 217 TrackHubs for upload to the UCSC Genome Browser. All custom scripts for data processing and 218 analysis are available at https://github.com/NordNeurogenomicsLab/. 219 220 RESULTS 221 222 Consistent patterns of transcriptional pathology associated with CHD8 haploinsufficiency 223 224 We reanalyzed a total of 240 RNA sequencing libraries corresponding to 10 studies of 225 CHD8 knockdown or heterozygous mutation (Table 1). Almost all datasets represented neuronal 226 model systems except for one dataset using an acute myeloid leukemia cell line (Shen et al.
6 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes
227 2015). Analysis of all datasets was performed using the same pipeline with quality control steps 228 and study-specific exceptions for consistency and covariate and batch structure as described in 229 original publication (Figure 1A). Unsurprisingly, relative gene expression levels varied widely 230 across studies, with principle components of variation dominated by species of origin and 231 experiment (Figure 1B). However, pairwise comparisons between DEGs from individual 232 datasets revealed specific similarities in gene expression changes. For example, comparison of 233 DEGs at the p < 0.01 cutoff level between the Gompers et al. (2017) and the Sugathan et al. 234 (2014) datasets revealed a strong positive correlation in direction of differential gene expression, 235 where genes that were significantly up- or down-regulated in one dataset followed the same 236 pattern in the other (Figure 1C). Further pairwise comparisons between studies and expression 237 for specific genes can be done using our interactive web browser available at 238 https://nordlab.shinyapps.io/rna_browser/. This interactive resource allows for analysis of 239 principle components, differential expression of individual genes, and overall differential 240 expression patterns for all included datasets (Supplemental Figure 1). New data from CHD8 241 models will be added to this site as they are published and available. 242 243 Considering expression of CHD8 itself, most knockdown and heterozygous knockout 244 models resulted in a 50-60% significant decrease in mRNA (Figure 1D). However, published 245 data from some models only showed a subtle decrease or even a significant increase in CHD8. 246 We verified that these findings were consistent with originally published RNA-seq data. The 247 absence of reduced CHD8 mRNA expression for some studies raises questions regarding what 248 expectations should be for gene dosage models. We note that protein level validation of CHD8 249 dosage decrease was performed in all original publications to confirm CHD8 haploinsufficiency 250 in each model but considering the use of a number of different and unvalidated CHD8 antibodies 251 across the studies, it is impossible to compare the protein validation results. 252 253 As expected, across all studies there were upregulated and downregulated genes passing 254 stringent thresholds, though the numbers of DEGs varied widely. Consistent with the original 255 publications, this re-analysis demonstrates CHD8 has direct or indirect roles in both facilitating 256 and repressing gene expression (Figure 1E). Large differences in number and effect size of 257 differentially expressed genes across studies may be a result of differences in experimental 258 design, impact of knockdown and knockout on CHD8 dosage, methods, and statistical sensitivity 259 related to intra-study sample variability and sequencing depth. Variability in gene expression 260 could also be due to differences in sensitivity to CHD8 dosage between developmental stages 261 and type of model used to carry out these experiments. 262 263 Changes in CHD8 expression affect genes important for cellular homeostasis and neuronal 264 function 265 266 To examine patterns of transcriptional dysregulation associated with knockdown or 267 heterozygous mutation to CHD8, we performed gene set enrichment analysis of Gene Ontology 268 (GO) terms including datasets with at least 500 differentially expressed genes (Figure 2). 269 Specific GO terms were chosen based on observed changes or interest to the field based on 270 previous findings (for example, “canonical Wnt signaling pathway”). The full list of GO terms 271 and relative enrichment is also provided (Supplementary Figure 2). While relatively small 272 numbers of individual genes showed overlapping significant changes in expression across
7 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes
273 pairwise study comparisons, we found strong correlation in DEG functional groups across 274 studies at the gene set level. This analysis identified two general signatures of differential gene 275 expression across published models. The first signature was characterized by downregulation of 276 genes annotated to terms related to the regulation of chromatin, transcription, and RNA 277 processing, which we refer to as general cellular homeostasis (homeostasis) genes. As a whole, 278 these are genes that do not exhibit cell specificity and are necessary for basic cell functions, such 279 as chromatin organization, transcription and translation, and mitosis. This includes terms such as 280 “RNA splicing,” “regulation of gene expression,” and “cell cycle.” The second signature 281 encompassed terms related to neuronal development, maturation, and function, including terms 282 associated with neural progenitor activity and lineage specification, synaptic function, and cell 283 adhesion. We refer to these genes as neuronal function (neuronal) genes, and these genes showed 284 both down- and up-regulation depending on the model. Examples of these terms include “neuron 285 differentiation,” “axon guidance,” and “cell adhesion.” 286 287 GO analysis showed different patterns of expression between upregulated and 288 downregulated genes. Upregulated genes were mainly enriched for neuronal terms across models 289 (Figure 2, top). In contrast, downregulated genes were enriched in homeostatic and neuronal 290 terms in distinct patterns (Figure 2, bottom). Of 22 datasets, around 7 had neuronal terms 291 enriched, 7 had homeostatic terms enriched, and 8 had a combination of both. The trend of 292 enrichment of these signatures showed some correlation to the model system used in each study. 293 In vitro models were more likely to have neuronal terms represented while in vivo models were 294 more likely to have both, or only homeostatic terms, represented. There is also some indication 295 that in vivo models of postnatal brain were more likely to have enrichment of neuronal terms 296 while models of embryonic brain were more likely to have enrichment of homeostasis terms, but 297 this remains a preliminary assessment requiring more robust data across developmental stages. 298 Hierarchical clustering of all RNA-seq datasets reinforced this pattern, though enrichment of 299 these signatures was weaker for datasets with fewer than 500 differentially expressed genes 300 (Supplementary Figure 3). We note that there were also GO terms enriched only in individual 301 datasets (Supplementary Figure 2). Overall, our results suggest that CHD8 knockdown or 302 heterozygous knockout consistently influences homeostatic and neuronal pathways, which are 303 likely to drive the cellular, anatomical, and behavioral pathology reported in studies of CHD8 304 haploinsufficiency. 305 306 CHD8-DNA interactions occur throughout the genome enriched for promoters 307 308 We reanalyzed a total of 49 ChIP-seq sequencing libraries from 8 studies of CHD8 309 genomic interaction patterns (Table 2). Analyzed datasets represented both neuronal and non- 310 neuronal model systems. We included both in vivo tissue preparations and in vitro culture models 311 from neuronal and non-neuronal fate cells to allow additional examination of tissue or cell-type 312 specificity of CHD8 interactions. Half of the datasets were generated from bulk mouse tissue at 313 adult (3 studies; Gompers et al. 2017, Katayama et al. 2016, Platt et al. 2017) and embryonic (2 314 studies; Katayama et al. 2016, Cotney et al. 2015) timepoints allowing for investigation of CHD8 315 interactions in vivo across time. The remaining data were generated from cellular models, with 316 two studies using human neuronal lineage cells (Sugathan et al. 2014, Cotney et al. 2015), two 317 using mouse or human cancer cell lines (Ceballos-Chavez et al. 2015, Shen et al. 2015), and one 318 using mouse embryonic stem cells (de Dieuleveult et al. 2016). ChIP-seq data were analyzed
8 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes
319 using the same steps for immunoprecipitated, or experimental, and control data in our analysis 320 pipeline (Figure 3A). There was large variation in number of called peaks, likely due to 321 experimental design and technical differences (Figure 3B). Eleven of the control ChIP-seq 322 libraries were found to have more than 250 called peaks with strong promoter enrichment 323 (Figure 3B-C), suggesting some level of technical artifact associated with chromatin preparation 324 (Marinov et al. 2014). Considering these experimental issues, control ChIP-seq libraries with 325 >250 peaks were included in the analysis to test for similarity between technical artifacts and 326 CHD8 immunoprecipitated signatures in these datasets. 327 328 Across all ChIP-seq datasets, CHD8 genomic interactions most commonly occurred near 329 promoters (Figure 3C). Furthermore, binding to promoter-defined peaks tended to approach 330 100% as the number of called peaks decreased, suggesting that higher affinity interactions for 331 CHD8 are largely at promoters. Increased affinity and frequency of promoter interactions by 332 CHD8 was clearly evident in the coverage data signal for both mouse tissues (Figure 4A) and 333 human cell lines (Figure 4B). For example, four genes encoding a transcription factor (ADNP), a 334 chromatin remodeler (SUV420H1), a splicing factor (TRA2B), and a calcium binding protein 335 important for cell cycle progression and ion channel signaling (CALM2) displayed CHD8 336 interactions almost exclusively near promoters. 337 338 De novo motif analysis performed on CHD8 peak regions across experiments identified 339 various general promoter-associated transcription factor binding sequences, but no clear primary 340 binding motif for CHD8 (Supplementary Figure 4). These findings are consistent with original 341 publications, none of which identified a strong candidate primary binding motif, suggesting that 342 CHD8 interactions are not mediated by direct DNA sequence recognition. Instead these results 343 suggest that CHD8 genomic interaction specificity likely occurs through secondary interactions. 344 345 CHD8 GO analysis of the top 2000 called peaks ranked by signal strength and 346 significance highlighted surprisingly consistent CHD8 interactions near homeostatic gene 347 promoters at the FDR < 0.05 GO term association cutoff level (Figure 5 top left). When 348 analyzing all called peaks, these terms were still enriched (Figure 5 bottom left). Analysis of all 349 called peaks also identified CHD8 interactions with the expanded set of promoters that included 350 genes associated with neuronal differentiation and function, as previously observed (Figure 4). 351 However, unlike homeostatic gene targets, neuronal gene interaction was not statistically 352 enriched across all relevant models suggesting the neuronal promoter interactions tended to have 353 lower affinity and are not as a set enriched as CHD8 regulatory targets. Testing for the 354 proportion of genes enriched for each GO term further indicated homeostatic gene promoter 355 target specificity. We found significant enrichment of homeostatic promoter targets when 356 analyzing the top 2000 peaks (Figure 5 top right). In contrast, almost all other GO terms were 357 represented when considering all called peaks (Figure 5 bottom right). This suggested that 358 while CHD8 has interactions throughout the genome with loci associated with various functions, 359 homeostatic genes are consistent, high affinity CHD8 targets regardless of the model system. 360 361 362 363
9 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes
364 Relationship between genomic interaction targets and gene expression changes across 365 CHD8 studies 366 367 Given that CHD8 targets a consistent set of promoters with a high level of affinity as well 368 as an expanded set of loci with lower levels of affinity and CHD8 knockdown or heterozygous 369 knockout causes changes to gene expression, we tested whether CHD8 genomic interactions 370 directly relate to changes in gene expression in CHD8 models. Most genes with CHD8 371 interactions at or distal to the promoter did not exhibit significant changes in gene expression, 372 regardless of the study, suggesting that there are additional determinants regarding sensitivity of 373 regulatory target genes to CHD8 dosage. While we did observe consistent patterns of overlap 374 between CHD8 targets and downregulated DEGs, upregulated DEGs were not enriched across 375 studies for CHD8 genomic interactions. Regardless of the experiment, CHD8 interaction affinity 376 was also strongest for genes that were more highly expressed (Supplementary Figure 5). 377 378 For a subset of RNA-seq results, there was a strong overlap between CHD8 target genes 379 from the ChIP-seq data and downregulated DEGs involved with cellular homeostasis. For 380 example, when using the Gompers et al. (2017) RNA-seq data, there was an increased signature 381 of downregulation in Chd8 heterozygous mouse brain as CHD8 target affinity (i.e. ChIP-seq 382 peak strength) increased (Figure 6). This was not unexpected considering homeostatic genes 383 made up one of the two major signatures of differential gene expression we found and was the 384 most strongly enriched set for CHD8 interaction. Eight out of the 18 analyzed datasets showed 385 this overlap trend (Supplementary Figure 6). For instance, an increased signature of 386 downregulation was also observed as CHD8 target affinity increased with in vivo Chd8 387 knockdown in fetal mouse brain (Durak et al. 2016), though this RNA dataset had both neuronal 388 and homeostatic terms enriched for DEGs (Supplementary Figure 7). 389 390 Genes associated with cellular homeostasis also tend to be at the high end of transcript 391 expression level distributions, suggesting a relationship between highly expressed genes and 392 dosage-sensitive CHD8 regulatory function. However, high levels of expression alone did not 393 predict CHD8 interaction or DEG, indicating that expression level does not solely determine 394 CHD8 interactions or sensitivity of regulatory targets to reduced CHD8 dosage. This trend of 395 negative correlation between CHD8 interaction affinity and changes in gene expression was less 396 apparent with datasets having fewer than 500 differentially expressed genes and for the datasets 397 (including many generated from in vitro models) where homeostasis gene expression signatures 398 were not present (Supplementary Figure 6, Figure 2). 399 400 While our findings show model-specific variation, the patterns present across CHD8 401 studies suggest a consistent relationship where reduced expression of CHD8 leads to 402 downregulation of CHD8 target genes associated with the cellular homeostasis signature, such as 403 genes involved in cell cycle, chromatin organization, and RNA transcription and processing. 404 These changes are seemingly stronger in in vivo models representing early stages of brain 405 development, though they are still present in some models representing more mature brain tissue 406 and cell types (Figure 2). In contrast, the observed differential expression of neuronal 407 differentiation and neuronal function (e.g. synaptic) genes tends to occur in models representing 408 more mature neuronal tissue or cell types and differentiated culture models inherently containing 409 heterogenous cell populations at unknown stages of development.
10 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes
410 DISCUSSION 411 412 This meta-analysis of published genomic datasets from in vitro and in vivo mouse and 413 human studies revealed both consistent and study-specific effects of CHD8 haploinsufficiency 414 on gene expression and largely concordant high-affinity CHD8 genomic interaction loci. 415 Knockdown or heterozygous mutation of CHD8 led to characteristic changes in gene expression 416 across studies and model systems. At the gene-by-gene level, these expression changes varied 417 considerably between CHD8 models. However, at the level of gene set enrichment, we found 418 global patterns of transcriptional dysregulation of genes involved in cellular homeostasis and 419 neuronal development and function. Comparison across ChIP-seq experiments shows that CHD8 420 preferentially targets promoters, with no evidence of direct binding through a specific DNA 421 motif. Surprisingly, we found that peaks with the highest signal were constant across 422 experiments, regardless of the model, suggesting that CHD8 preferentially interacts with 423 promoters of a set of genes linked to processes involved in cellular homeostasis, genome 424 function, and RNA processing. Our findings strongly support signatures of reduced transcription 425 of CHD8 target genes in the studied models that were dependent on dosage, though our data also 426 highlight widespread genomic promoter interactions for CHD8 without obvious strong impacts 427 to most targets. We verified the presence of changes to gene expression specific to neuronal 428 differentiation and function following CHD8 haploinsufficiency across studies, however, these 429 changes do not appear to be through direct disruption to neural cell-type or stage-specific CHD8 430 regulatory activity via high-affinity interactions with relevant promoters. While the clear 431 concordance in high-affinity genomic CHD8 interactions suggests common regulatory functions 432 across cell types, it remains to be examined whether the observed dysregulation of neuronal 433 genes is related to context-dependent CHD8 regulatory activity in the brain given the current 434 cellular heterogeneity and technical challenges existing with available CHD8 ChIP-seq. Our 435 results illustrate the power and limitations of comparing genomic datasets and challenge previous 436 assumptions regarding the regulatory mechanisms and transcriptional pathology associated with 437 CHD8 haploinsufficiency. 438 439 We note a number of technical issues that impacted this meta-analysis, many of which 440 are associated with variation in methods and sequencing depth. Surprisingly, we found 441 considerable differences in CHD8 expression across models despite the common design of 442 testing the impacts of haploinsufficiency. Though we did not find an obvious correlation between 443 CHD8 transcript levels and up- or downregulated gene expression, it seems likely that 444 differences in experimental design, including CHD8 knockdown or knockout, contributed toward 445 meaningful variation between models. Changes to CHD8 dosage have been shown to have strong 446 and potentially opposing effects on cellular function. For instance, homozygous knockout of 447 CHD8 has been described to cause severe developmental arrest and widespread apoptosis 448 leading to early embryonic lethality (Nishiyama et al. 2009) while heterozygous mutation can 449 lead to increases in proliferation (Gompers et al. 2017). These are important considerations for 450 interpreting studies of haploinsufficiency, as allelic or genetic background effects as well as 451 variation in transcriptional knockdown with shRNA constructs may have significant biological 452 consequences. At a minimum, consistent measures of CHD8 knockdown or haploinsufficiency, 453 such as measuring transcript-level mRNA levels with RNA-seq or protein levels with a 454 standardized antibody and methodology, should be a goal for future studies to enable comparison 455 across publications. We also noted differences in ChIP-seq datasets that hindered comparisons.
11 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes
456 For example, enrichment in control libraries was present across several published datasets. 457 Different studies also used various CHD8 antibodies with unknown and unvalidated CHD8 458 specificities. Nonetheless, by examining patterns across datasets, we identified consistent 459 patterns of enrichment suggesting that overall findings from ChIP-seq targeting CHD8 could 460 reliably identify high affinity interactions. 461 462 Despite the limitations of comparing genomic datasets across variable models, our 463 analysis challenges two simple models regarding pathological mechanisms of CHD8 464 haploinsufficiency. The first model the transcriptional signatures present across studies refute is 465 that pathology due to CHD8 haploinsufficiency is primarily due to alterations in patterning 466 during early brain development. While our meta-analysis clearly supports impacts to 467 proliferation and neuronal differentiation consistent with published findings on proliferation and 468 brain volume (Bernier et al. 2014, Gompers et al. 2017, Katayama et al. 2016, Platt et al. 2017, 469 Durak et al. 2016), we also observed evidence of dysregulation of genes involved in mature 470 neuron function, including synaptic genes. This is consistent with observation that CHD8 is still 471 highly expressed in adulthood (Gompers et al. 2017, Platt et al. 2017, Maussion et al. 2015), that 472 mutations to CHD8 continue to lead to differential gene expression and behavioral phenotypes in 473 adult mice (Gompers et al. 2017, Katayama et al. 2016, Platt et al. 2017), and with limited 474 evidence of synaptic dysfunction associated with Chd8 haploinsufficiency (Platt et al. 2016). 475 Further work will be required to establish the role and requirement for CHD8 in mature neurons 476 and other cell types in the brain. 477 478 Second, the signatures present in this meta-analysis suggest that pathology observed in 479 CHD8 models and patients with CHD8 mutations is not due to targeted impacts to specific 480 populations of cell-types or due to impacts limited to specific brain regions. In this analysis of 481 many individual datasets, CHD8 had genomic interactions near promoters of genes important for 482 cellular homeostasis and neuronal development and function that were enriched in the 483 transcriptomic analysis. However, only homeostasis genes were characterized as high affinity 484 CHD8 targets and tended to be sensitive to decreases in CHD8 expression and 485 haploinsufficiency. Despite evidence that these genes are not high-affinity CHD8 targets, we did 486 observe enrichment of differentially expressed neuronal genes in the CHD8 interaction analysis. 487 One explanation for this finding is that CHD8 haploinsufficiency indirectly causes large-scale 488 dysregulation of neuronal genes via disruptions to upstream transcriptional regulators that are 489 direct CHD8 targets. This would explain the appearance of cell-type-specific transcriptional 490 changes in the absence of actual cell-type specific CHD8 function. Nonetheless, given the 491 technical limitations of current studies we cannot rule out the possibility of cell-type or context- 492 dependent specificity of CHD8 function. 493 494 It is clear from previous publications and this meta-analysis that CHD8 is critical for 495 neurodevelopment. Our results suggest that CHD8 functions to regulate cellular homeostasis 496 required for genomic control of proliferation and differentiation. As an essential gene with 497 widespread expression across neuronal and glial cell types, homozygous loss of CHD8 may 498 impact cellular viability in general, while heterozygous mutation or knockdown might have 499 subtler, context-specific impacts. Such a model would explain the widespread changes in gene 500 expression across model systems and varied reports of impact on proliferation depending on 501 dosage. Our results raise two questions that could be addressed by application of RNA-seq and
12 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes
502 ChIP-seq in the future: 1) What are the developmental stage, cell-type, and region-specific 503 impacts of CHD8 haploinsufficiency in the developing and mature brain, and 2) Does CHD8 504 have context-dependent function in specific stages, cell types, and regions with regard to 505 genomic interaction patterns? Beyond addressing these two key issues, additional clarity 506 regarding the role of CHD8 in the brain will come from studies examining the molecular 507 interaction partners and impacts on chromatin, transcription, and RNA processing. As CHD8 508 haploinsufficiency may represent common features of haploinsufficiency of other general 509 chromatin remodelers implicated in patient studies, further characterization of CHD8 models and 510 CHD8 genomic interactions could reveal essential functions driving pathology in 511 neurodevelopmental disorders. 512 513 AUTHOR CONTRIBUTIONS 514 515 AW and AN conceived of the project. AW, KL, and AN performed analysis of RNA-seq 516 experiments. AW, RCP, and AN performed analysis of ChIP-seq experiments. AW and AN 517 drafted the manuscript. All authors contributed to manuscript revision. 518 519 FUNDING 520 521 AW was supported by Training Grant number T32-GM007377 from NIH-NIGMS. RCP was 522 supported by a Science Without Borders Fellowship from CNPq (Brazil). AN was supported by 523 NIH-NIGMS R35 GM119831. 524 525 ACKNOWLEDGEMENTS 526 527 The authors would like to thank the authors from the original CHD8 studies who provided access 528 to the raw data for analysis. 529 530 CONFLICT OF INTEREST STATEMENT 531 532 The authors declare that there is no conflict of interest.
13 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes
REFERENCES
Barnard, R. A., Pomaville, M. B., and O’Roak, B. J. (2015). Mutations and modeling of the chromatin remodeler CHD8 define an emerging autism etiology. Front. Neurosci. 9, 477. doi: 10.3389/fnins.2015.00477 Bernier, R., Golzio, C., Xiong, B., Stessman, H., Coe, B., Penn, et al. (2014). Disruptive CHD8 mutations define a subtype of autism early in development. Cell 158, 263-276. doi: 10.1016/j.cell.2014.06.017 Ceballos-Chávez, M., Subtil-Rodríguez, A., Giannopoulou, E., Soronellas, D., Vázquez-Chávez, E., Vicent, G., et al. (2015). The chromatin remodeler CHD8 is required for activation of progesterone receptor-dependent enhancers. PLoS Genet. 11, e1005174. doi: 10.1371/journal.pgen.1005174 Cotney, J., Muhle, R., Sanders, S., Liu, L., Willsey, A., Niu, W., et al. (2015). The autism- associated chromatin modifier CHD8 regulates other autism risk genes during human neurodevelopment. Nat. Comm. 6, 6404. doi: 10.1038/ncomms7404 de Dieuleveult, M., Yen, K., Hmitou, I., Depaux, A., Boussouar, F., Dargham, D., et al. (2016). Genome-wide nucleosome specificity and function of chromatin remodellers in ES cells. Nature 530, 113-116. doi: 10.1038/nature16505 De Rubeis, S., He, X., Goldberg, A., Poultney, C., Samocha, K., Ercument Cicek, A., et al. (2014). Synaptic, transcriptional and chromatin genes disrupted in autism. Nature 515, 209-215. doi: 10.1038/nature13772 Dobin, A., Davis, C. A., Schlesinger, F., Drenkow, J., Zaleski, C., Jha, S., et al. (2013). STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21. doi: 10.1093/bioinformatics/bts635 Durak, O., Gao, F., Kaeser-Woo, Y., Rueda, R., Martorell, A., Nott, A., et al. (2016). Chd8 mediates cortical neurogenesis via transcriptional regulation of cell cycle and Wnt signaling. Nat. Neurosci. 19, 1477-1488. doi: 10.1038/nn.4400 Fang, M., Hutchinson, L., Deng, A., and Green, M. R. (2016). Common BRAF(V600E)-directed pathway mediates widespread epigenetic silencing in colorectal cancer and melanoma. PNAS 113, 1250–1255. doi: 10.1073/pnas.1525619113 Feng, J., Liu, T., and Zhang, Y. (2011). Using MACS to identify peaks from ChIP-Seq data. Curr. Protoc. Bioinformatics Chapter 2, Unit 2.14. doi: 10.1002/0471250953.bi0214s34 Flanagan, J., Mi, L., Chruszcz, M., Cymborowski, M., Clines, K., Kim, Y., et al. (2005). Double chromodomains cooperate to recognize the methylated histone H3 tail. Nature 438, 1181-1185. doi: 10.1038/nature04290 Gompers, A.L., Su-Feher, L., Ellegood, J., Copping, N.A., Riyadh, A., Stradleigh, T.W., et al. (2017). Germline Chd8 haploinsufficiency alters brain development in mouse. Nat. Neurosci. 20, 1062–1073. doi: 10.1038/nn.4592 Hall, J., and Georgel, P. (2007). CHD proteins: a diverse family with strong ties. Biochem. Cell Biol. 85, 463-476. doi: 10.1139/O07-063 Han, H., Braunschweig, U., Gonatopoulos-Purnatzis, T., Weatheritt, R.J., Hirsch, C.L., Ha, K.C., et al. (2017). Multilayered Control of Alternative Splicing Regulatory Networks by Transcription Factors. Mol. Cell 65, 539-553. doi: 10.1016/j.molcel.2017.01.011
14 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes
Hargreaves, D., and Crabtree, G. (2011). ATP-dependent chromatin remodeling: genetics, genomics and mechanisms. Cell Res. 21, 396-420. doi: 10.1038/cr.2011.32 Heinz, S., Benner, C., Spann, N., Bertolino, E., Lin, Y. C., Laslo, P., et al. (2010). Simple combinations of lineage-determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities. Mol. Cell 38, 576–589. doi: 10.1016/j.molcel.2010.05.004 Iossifov, I., O’Roak, B. J., Sanders, S. J., Ronemus, M., Krumm, N., Levy, D., et al. (2014). The contribution of de novo coding mutations to autism spectrum disorder. Nature 515, 216–221. doi: 10.1038/nature13908 Ishihara, K., Oshimura, M., and Nakao, M. (2006). CTCF-dependent chromatin insulator is linked to epigenetic remodeling. Mol. Cell 23, 733-742. doi: 10.1016/j.molcel.2006.08.008 Katayama, Y., Nishiyama, M., Shoji, H., Ohkawa, Y., Kawamura, A., Sato, T., et al. (2016). CHD8 haploinsufficiency results in autistic-like phenotypes in mice. Nature 537, 675- 679. doi: 10.1038/nature19357 Krumm, N., O’Roak, B., Shendure, J., and Eichler, E. (2014). A de novo convergence of autism genetics and molecular neuroscience. Trend Neurosci. 37, 95-105. doi: 10.1016/j.tins.2013.11.005 Li, H., and Durbin, R. (2009). Fast and accurate short read alignment with Burrows-Wheeler transform. Bioinformatics 25, 1754–1760. doi: 10.1093/bioinformatics/btp324 Liao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics 30, 923–930. doi: 10.1093/bioinformatics/btt656 Marfella, C., and Imbalzano, A. (2007). The Chd family of chromatin remodelers. Mutat. Res. Fund. Mol. Mech. Mut. 618, 30-40. doi: 10.1016/j.mrfmmm.2006.07.012 Marinov, G. K., Kundaje, A., Park, P. J., and Wold, B. J. (2014). Large-scale quality analysis of published ChIP-seq data. G3 4, 209–223. doi: 10.1534/g3.113.008680 Maussion, G., Diallo, A.B., Gigek, C. O., Chen, E. S., Crapper, L., Théroux, J. F., et al. (2015). Investigation of genes important in neurodevelopment disorders in adult human brain. Hum. Genet. 134, 1037-1053. doi: 10.1007/s00439-015-1584-z McCarthy, S. E., Gillis, J., Kramer, M., Lihm, J., Yoon, S., Berstein, Y., et al. (2014). De novo mutations in schizophrenia implicate chromatin remodeling and support a genetic overlap with autism and intellectual disability. Mol. Psych. 19, 652–658. doi: 10.1038/mp.2014.29 McKnight, J., Jenkins, K., Nodelman, I., Escobar, T., and Bowman, G. (2011). Extranucleosomal DNA binding directs nucleosome sliding by Chd1. Mol. Cell Biol. 31, 4746-4759. doi: 10.1128/MCB.05735-11 Nishiyama, M., Oshikawa, K., Tsukada, Y., Nakagawa, T., Iemura, S., Natsume, T., et al. (2009). CHD8 suppresses p53-mediated apoptosis through histone H1 recruitment during early embryogenesis. Nat. Cell Biol. 11, 172–182. doi: 10.1038/ncb1831 O'Roak, B., Vives, L., Fu, W., Egertson, J., Stanaway, I., Phelps, I., et al. (2012a). Multiplex targeted sequencing identifies recurrently mutated genes in autism spectrum disorders. Science 338, 1619-1622. doi: 10.1126/science.1227764
15 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes
O’Roak, B., Vives, L., Girirajan, S., Karakoc, E., Krumm, N., Coe, B., et al. (2012b). Sporadic autism exomes reveal a highly interconnected protein network of de novo mutations. Nature 485, 246-250. doi: 10.1038/nature10989 Parikshak, N., Luo, R., Zhang, A., Won, H., Lowe, J., Chandran, V., et al. (2013). Integrative functional genomic analyses implicate specific molecular pathways and circuits in autism. Cell 155, 1008-1021. doi: 10.1016/j.cell.2013.10.031 Platt, R., Zhou, Y., Slaymaker, I., Shetty, A., Weisbach, N., Kim, J., et al. (2017). Chd8 mutation leads to autistic-like behaviors and impaired striatal circuits. Cell Rep. 19, 335-350. doi: 10.1016/j.celrep.2017.03.052 Robinson, M. D., McCarthy, D. J., and Smyth, G. K. (2010). edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26, 139– 140. doi: 10.1093/bioinformatics/btp616 Rodriguez-Paredes, M., Ceballos-Chavez, M., Esteller, M., Garcia-Dominguez, M., and Reyes, J. (2009). The chromatin remodeling factor CHD8 interacts with elongating RNA polymerase II and controls expression of the cyclin E2 gene. Nucl. Acid Res. 37, 2449- 2460. doi: 10.1093/nar/gkp101 Sanders, S. J., He, X., Willsey, A. J., Ercan-Sencicek, A. G., Samocha, K. E., Cicek, A. E., et al. (2015). Insights into autism spectrum disorder genomic architecture and biology from 71 risk loci. Neuron 87, 1215–1233. doi: 10.1016/j.neuron.2015.09.016 Shen, C., Ipsaro, J. J., Shi, J., Milazzo, J. A., Wang, E., Roe, J.-S., et al. (2015). NSD3-short is an adaptor protein that couples BRD4 to the CHD8 chromatin remodeler. Mol. Cell 60, 847–859. doi: 10.1016/j.molcel.2015.10.033 Sugathan, A., Biagioli, M., Golzio, C., Erdin, S., Blumenthal, I., Manavalan, P., et al. (2014). CHD8 regulates neurodevelopmental pathways associated with autism spectrum disorder in neural progenitors. P.N.A.S. 111, E4468-E4477. doi: 10.1073/pnas.1405266111 Tatton-Brown, K., Loveday, C., Yost, S., Clarke, M., Ramsay, E., Zachariou, A., et al. (2017). Mutations in epigenetic regulation genes are a major cause of overgrowth with intellectual disability. Am. J. Hum. Genet. 100, 725–736. doi: 10.1016/j.ajhg.2017.03.010 Thompson, B., Tremblay, V., Lin, G., and Bochar, D. (2008). CHD8 is an ATP-Dependent chromatin remodeling factor that regulates beta-catenin target genes. Mol. Cell Biol. 28, 3894-3904. doi: 10.1128/MCB.00322-08 Tong, J.K., Hassig, C.A., Schnitzler, G.R., Kingston, R.E., and Schreiber, S.L. (1998). Chromatin deacetylation by an ATP-dependent nucleosome remodeling complex. Nature 395, 917-921. doi: 10.1038/27699 Vissers, L. E. L. M., Gilissen, C., and Veltman, J. A. (2016). Genetic studies in intellectual disability and related disorders. Nat. Rev. Genet. 17, 9–18. doi: 10.1038/nrg3999 Wang, L., Wang, S., and Li, W. (2012). RSeQC: quality control of RNA-seq experiments. Bioinformatics 28, 2184–2185. doi: 10.1093/bioinformatics/bts356 Wang, P., Lin, M., Pedrosa, E., Hrabovsky, A., Zhang, Z., Guo, W., et al. (2015). CRISPR/Cas9- mediated heterozygous knockout of the autism gene CHD8 and characterization of its transcriptional networks in neurodevelopment. Mol. Autism 6, 55. doi: 10.1186/s13229-015-0048-6
16 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes
Wang, P., Mokhtari, R., Pedrosa, E., Kirschenbaum, M., Bayrak, C., Zheng, D., et al. (2017). CRISPR/Cas9-mediated heterozygous knockout of the autism gene CHD8 and characterization of its transcriptional networks in cerebral organoids derived from iPS cells. Mol. Autism 8, 11. doi: 10.1186/s13229-017-0124-1 Wilkinson, B., Grepo, N., Thompson, B. L., Kim, J., Wang, K., Evgrafov, O. V., et al. (2015). The autism-associated gene chromodomain helicase DNA-binding protein 8 (CHD8) regulates noncoding RNAs and autism-related genes. Transl. Psych. 5, e568. doi: 10.1038/tp.2015.62 Young, M. D., Wakefield, M. J., Smyth, G. K., and Oshlack, A. (2010). Gene ontology analysis for RNA-seq: accounting for selection bias. Genome Biol. 11, R14. doi: 10.1186/gb- 2010-11-2-r14 Yuan, C., Zhao, X., Florens, L., Swanson, S., Washburn, M., and Hernandez, N. (2007). CHD8 associates with human Staf and contributes to efficient U6 RNA polymerase III transcription. Mol. Cell Biol. 27, 8729-8738. doi: 10.1128/MCB.00846-07
17 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes
FIGURES
Table 1
Summary of RNA-seq datasets included in the CHD8 model reanalysis; h – human, NSCs – neural stem cells, iPSC – induced pluripotent stem cell, NPC – neural progenitor cell, E – embryonic day, P – postnatal day, Ctx – Prefrontal Cortex, Striat – Dorsal Striatum, Nuc Acc – Nucleus Accumbens, VTA – Ventral Tegmental Area, Hipp – Hippocampal Formation, Amyg – Amygdala, Hyp – Lateral Hypothalamus.
Manipulation Study Model System Timepoint(s) Tissue Type
Cotney et al. shRNA H9-derived hNSCs - - 2015 transfection
Durak et al. E13 shRNA GFP+ Cortical Swiss Webster mice E15 2016 electroporation Cells
CHD8 Sugathan et al. shRNA iPSC-derived hNPCs - - Knockdown 2014 transfection
Wilkinson et al. SK-N-SH siRNA - - 2015 hNeuroblastoma cell transfection
MLL-AF9/NrasG12D shRNA Shen et al. 2015 - - mAML (RN2) cells transfection
Gompers et al. CRISPR-cas9; E: 12.5, 14.5, 17.5 C57BL/6N mice Bulk Forebrain 2017 Exon 5 (5bp) P: 0, 56 Cre-LoxP Recombination Katayama et al. (ESC clones E: 10, 12, 14, 16, 18 C57BL/6J mice Whole Brain 2016 injected into P: 91 blastocysts); Exon 11-13 Ctx, Striat, CRISPR-cas9; Nuc Acc, Heterozygous Platt et al. 2017 C57BL/6J P70-84 CHD8 Mutation Exon 1 (7bp) VTA, Hipp, Amyg, Hyp CRISPR-cas9; MLL-AF9/NrasG12D Shen et al. 2015 ATPase, Helicase - - mAML (RN2) cells domains
Wang et al. iPSC-derived hNPCs, CRISPR-cas9, N- - - 2015 hNeurons terminus
Wang et al. iPSC-derived CRISPR-cas9, N- - - 2017 hCerebral Organoids terminus
18 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes
Figure 1 Figure 1
A B
Mouse
Human
Cotney Katayama Sugathan Durak Platt Wang Gompers Shen Wilkinson C
-0.5
0
0.5
D Change Gompers Fold Log -3 -2 -1 0 1 2 FC Sugathan Log Fold Change 1.21.2 * * * 1.01.0 0.80.8 * * * * C * * * * * F 0.60.6 * * Fold Fold Change * * * * 0.40.4 * *
CHD8 0.20.2
0.00 DE.P05.count l l l l l l l l l l l l x c p p g 1 2 2 4 8 1 6 0 5 5 5 0 6 n d 8 8 8 A C C t G i m . . . . . u c u u u u . u y y . 1 1 1 9 1 1 5 o p T P D D D i o u 8 8 2 4 7 P
r
E C 8 F F F F F F 8 t A H E E E P E E _ P V m n Up 1 1 1 Up u H H H _ H N D D _ c _ _ _ _ _ a s D t ______D _ t i a A e t r _ t E E E t k s y _ t u a n C C C r H H s t t a a a a a a t i r H g H t _ a e r t _ _ _ Down e a g a a N h g r t l
a N s Down a l r 8000 C C m e t l s s s
8000 l a m m m m m m p e C S
n C n s s h _ l _ r r r _ t P u h h a p t O P a t a a a a a a p h _ P h a _ _ P g l t t e e e m n t y o P s s a _ y y y y y y s D s t 2 2 m n n P a p p p o m o a _ _ g g W a a a a a a C l _ _ t a o a t t t t t t s N N u l k k o u n y G y m m m a P
a a a a a a n o a a G P a R R e i W e G S o o o r r 60006000 K K K K K K K k _ _ C n n u u l
t W t G G G i n n e o o D D e e W n C C h h e S S
G 40004000
E D 2000 # of DE Genes 2000
0 0 l l l l l l l l l l l l x c p p g 1 2 2 4 8 1 6 0 5 5 5 0 6 n d 8 8 8 A C C t G i m . . . . . u c u u u u . u y y . 1 1 1 9 1 1 5 o p T P D D D
i o u 8 8 2 4 7 P r C 8 F F F F F F 8 t A H E E E P E E _ P V m n 1 1 1 u H H H _ H N D D _ c _ _ _ _ _ a s D t ______D _ t i a A e t r _ t E E E t k s y _ t u a n C C C r H H s t t a a a a a a t i r H g H t _ a e r t _ _ _ e a g a a N h g r t l a N s a l r C C m e t l s s s
l a m m m m m m p e C S n C n s s h _ l _ r r r _ t P u h h a p t O P a t a a a a a a p h _ P h a _ _ P g l t e e e m n t y o P s s a _ y y y y y y s D s t 2 2 m n P a p p p o m o a _ _ g g W a a a a a a C l _ _ t a o a t t t t t t s N N l k k o u n y G y m m m a
P a a a a a a n a a G P a R R e i W e G S o o o r r K K K K K K K k _ _ n n u u l t W t G G G i n n o o D D e e
W C C h h S S
Differential gene expression across CHD8 models. (A) RNA-seq data analysis pipeline. (B) PCA showing similarity in gene expression according to species. (C) Correlation between the Gompers et al. 2017 and Sugathan et al. 2014 RNA-seq datasets (p < 0.01) generated from the Shiny web browser. (D) Change in CHD8 mRNA across models (p < 0.05). (E) Differential expression genes count across models (p < 0.05).
19 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes
Figure 2
Enrichment of gene regulation and neurodevelopmental ontology terms in Up- and Down- regulated datasets having greater than 500 differentially expressed genes at the p < 0.05 level. (Top) Upregulated gene ontology enrichment. (Bottom) Downregulated gene ontology enrichment.
20 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes
Table 2
Summary of datasets included in CHD8 ChIP-seq reanalysis. All datasets were performed using formaldehyde or another similar method of crosslinking before fragmentation and immunoprecipitation; E – embryonic day, P – postnatal day, h – human, NSCs- Neural Stem Cells, m – mouse, ESCs – Embryonic Stem Cells, HA -haemagglutinin, MNase – Micrococcal nuclease, Chd8+/- – CHD8 heterozygous mutation carrier.
Fragmentation Tissue Study Model Timepoint(s) Antibody Control Method Collected
C57BL/6J Cotney et Frontal αCHD8 Sonication mice; H9- E17.5; - Input al. 2015 Cortex; - (Abcam, ab114126) derived hNSCs C57BL/6J Sonication & Katayama mice αCHD8 E14, P91 Whole Brain Input Mnase et al. 2016 (Chd8+/- & (Custom) WT) C57BL/6J αCHD8 Platt et al. mice Somatosensory Sonication P70-84 (Novus Biologicals, IgG 2017 (Chd8+/- & Cortex NB100-60417) WT) Ceballos- hT47D-MTVL Before αCHD8 Sonication Chavez et breast cancer progestin - IgG (Bethyl, A301-224A) al. 2015 cell stimulation de mESCs with MNase Dieulevuelt FLAG/HA- - - αFLAG & αHA Input et al. 2016 tagged CHD8
Gompers et C57BL/6N αCHD8 Sonication ~P56 Bulk Forebrain Input al. 2017 mice (Abcam, ab114126)
Shen et al. αCHD8 Sonication mRN2 cells - - Input 2015 (Bethyl, A301-224A) αCHD8 (Bethyl, A301-224A; Sugathan et iPSC-derived Sonication - - Novus Biologicals, Input al. 2014 hNPCs NB100-60417, NB100- 60418)
21 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes Figure 3 Figure 3
A
>250 Peaks
B Immunoprecipitated Control
40000 40000 3000030000 2000020000 # Peaks 1000010000 0 0
l t t t t t t t t t t 1 2 1 2 2 3 7 8 1 2 n 1 2 1 2 d
C T T T T D y G i p p p p p p p p p p ______1 1 ______A h 7 a H H g n n n n n n n n n n t W W
100 I r 2 2 4 4 7 7 I I I I I I I I I I
1.0 T T T T _ 4 C C C C _ _ e _ _ _ b 0 0 1 1 ______n
80 N N T H H d 4
0.8 S S S S i W W d 4 B e 6 6 1 2 2 3 d T T T T D E E C
_ _ _ 1 r R R A a
60 E E N N _ _ 1 s s
0.6 _ A _ _ _ _ r 7 _ _ A x x g H H S o _ _ _ E _ _ x x _ _ t t W W u u _ E 2 2
40 x x b t t 4 _ TSS % TSS
C 0.4 C C h _ _ t t n _ F P g g v v _ _ i n _ m m C C i e m m n T N N _ i d 4 S
20 C C S S l n i 0.2 _ o o h h d 4 C C
a i n _ _ _ z _ a i 1 _ _ r o m m R R A a P m m 1 _ _ N N
0 _ _ r a A i N N g
0.0 m r e g a b h r _ _ m m _ _ _ E _ _ y y
_ _ b r m m _ E _ _ _ h v b t t h l m m t t t t t t t t t t e b n n l l n _ m e e g g 1 2 1 2 2 3 7 8 1 2 n 1 2 1 2 d T T T T l e b i n _ g D y m m m m _ a G i e p p p p p p p p p p W _ C C l i m m e l e e u u n
______1 1 ______n n h h A h 7 m m a H H g o l e a i n h n n n n n n n n n n _ _ n h t t z t _ W W I o l a i r 2 2 4 4 7 7 I I I I I I I I I I _ _ s T T T T S S r o h h e e t t m m _ 4 C C C C _ _ Chavez _ _ Shen_1 Shen_2 _ _ h o r t t a _ e _ _ t t a r o o _ e - b 0 0 1 1 ______h o a C n v v t t N N T H H d 4 b h r _ _ Gompers S S S S y y i P P m S S y y W W d 4 h b r B n e v 6 6 1 2 2 3 d a a h e T T T T i i D E E C h _ _ _ 1 e e C C − l l a a r R R A a e b n n t E E e e W N N _ _ 1 e e s s l l _ A _ _ _ _ l l Platt_HT_1 Platt_HT_2 l e b r a 7 _ _ A a W x x p _ _ g H H Platt_WT_1 Platt_WT_2 S m W o _ _ _ E s Chavez_IgG l _ _ x x Shen_1_Inpt Shen_2_Inpt _ _ t t W W e e e a u u n n _ W _ E 2 2 P P x x n n u u - b t t 4 _ C C C o l e h _ _ h h _ W P P t t n _ g g t t F P t t _ _ g g v v _ _ Katayama_Ad o o l Cotney_Ctx_1 Cotney_Ctx_2 Sugathan_Inpt i n _ t m m C C i h h g m e m m n _ T N N l e e _ i d 4 S C C S S l n i h o _
Cotney_NSC_2 Cotney_NSC_3 _ o o h h d 4 l h h o o m i i C C o o a a i n C _ _ _ h o deDieuleveult_1 deDieuleveult_2 z _ a m u a i 1 o Ceballos _ _ Sugathan_Bethyl r o m m R R A a m S S P m m 1 _ _ N N h _ _ r a A i N N a _ _ g m m r e g − h a C C D D g C C b h r _ _ m m m _ _ _ E _ _ y y Katayama_Ad_HT m W _ _ m S b r m m _ E Katayama_Ad_WT _ _ _ h Katayama_Ad_Inpt G v b t t m W h b n n Katayama_E14_HT m m s m e b n n l l n _ m e e Katayama_E14_WT e e u g g Sugathan_Novus417 Sugathan_Novus418 _ m _ W l e b Cotney_NSC_2_Inpt Cotney_NSC_3_Inpt i n _ g a m m m m _ a e W _ C C l i _ m _ W m m e a a e l e e u u n _ n n o _ Ceballos h h m m d d o l e a i n h _ _ n h
y t t l a _ _ z _ S o l a i _ _ s S S r o h h e e t t m m a _ l _ _ _ h h m _ _ a h o r t t a Katayama_Ad_HT_Inpt _ t t a r o o C e m a Katayama_Ad_WT_Inpt h o a C t t a v v t t b h r _ _ Katayama_E14_HT_Inpt y y P P m S S y y a a t h b r Katayama_E14_WT_Inpt n m m v a a h e i i h e e C C − l l a a m m e b n n t e e a a W m e e l l m l l l e b a a W p b _ _ m a m W a m s l e e e a n n _ W P P n n a m u u g g _ m o l e h a h _ W P P g g t t t t _ _ y o o l e _ m a t h h g
m K _ l e e y h o a _ y u u a _ l h h o o m i i o o a C h o a a m u y o a _ m S S h C a y t a _ _ m a − h a C C D D g C C m S S t m W a m CHD8 binds to promoters across the genome. (A) ChIP-seq analysis pipeline. (B) S Number of a t m G m W b n n s a m a t e e m u _ m _ W a a t _ m _ W a e a a a m o _ d d a y l a _ _ a m S K a _ l called peaks meeting a MACS2 significance of p < 0.00001. (C) Percentage of called peaks _ a h h m a y K C a m a K t t a y a a a t K m m a y m m K a a m m a y b m a t overlapping with the transcription start site of the nearest gene.a m a m t g g _ m a a y e _ m a a K a t y a y u u a _ a t a y a _ C a y t a a a K S S t a a t a m K a a t m a t K a a m a a m K K a y K a K y a K a y K a y t t a a a t a t a K a K K K
22 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes
FigureFigure 4 4
A Adnp Suv420h1 Tra2b Calm2 10kb 20kb 10kb 10kb
Shen_mm_RN2 Dieu_mm_ESC Cotney_mm_Ctx Gompers_mm_Forebrain Platt_mm_Ctx_HT Platt_mm_Ctx_WT Katayama_mm_Wholebrain_Ad_HT Immunoprecipitated Katayama_mm_Wholebrain_Ad_WT Katayama_mm_Wholebrain_E14_HT Katayama_mm_Wholebrain_E14_WT Shen_mm_RN2_Inpt Katayama_mm_Wholebrain_Ad_HT_Inpt Katayama_mm_Wholebrain_Ad_WT_Inpt
Control Katayama_mm_Wholebrain_E14_HT_Inpt Katayama_mm_Wholebrain_E14_WT_Inpt
B ADNP SUV420H1 TRA2B CALM2 20kb 20kb 10kb 10kb
Chavez_hg_T427D Cotney_hg_NSC Sugathan_hg_NPC_Bethyl Sugathan_hg_NPC_Novus60417 Immunoprecipitated Sugathan_hg_NPC Novus60418 Chavez_hg_T427D_Inpt Cotney_hg_NSC_Inpt Control Sugathan_hg_NPC_Inpt
Examples of CHD8 binding near promoters of select chromatin (ADNP, SUV420H1), RNA processing (TRA2B), and neuronal function (CALM2) genes in the mouse (top) and human (bottom) ChIP-seq datasets.
23
D 7 4 T _ g h _ z e v a h C s o l l a b e C D 7 4 T _ g h _ z e v a h C s o l l a b e C − −
1 _ 2 N R _ m m _ n e h S 1 _ 2 N R _ m m _ n e h S
2 _ 2 N R _ m m _ n e h S 2 _ 2 N R _ m m _ n e h S
1 _ C S E _ m m _ t l u e v e l u e i D e d 1 _ C S E _ m m _ t l u e v e l u e i D e d
2 _ C S E _ m m _ t l u e v e l u e i D e d 2 _ C S E _ m m _ t l u e v e l u e i D e d
2 _ C S N _ g h _ y e n t o C 2 _ C S N _ g h _ y e n t o C
3 _ C S N _ g h _ y e n t o C 3 _ C S N _ g h _ y e n t o C
l y h t e B _ C S P i _ g h _ n a h t a g u S l y h t e B _ C S P i _ g h _ n a h t a g u S
7 1 4 0 6 s u v o N _ C S P i _ g h _ n a h t a g u S 7 1 4 0 6 s u v o N _ C S P i _ g h _ n a h t a g u S
8 1 4 0 6 s u v o N _ C S P i _ g h _ n a h t a g u S 8 1 4 0 6 s u v o N _ C S P i _ g h _ n a h t a g u S
1 _ 7 1 E _ _ x t C _ m m _ y e n t o C 1 _ 7 1 E _ _ x t C _ m m _ y e n t o C
2 _ 7 1 E _ x t C _ m m _ y e n t o C 2 _ 7 1 E _ x t C _ m m _ y e n t o C
n i a r b e r o F _ m m _ s r e p m o G n i a r b e r o F _ m m _ s r e p m o G
1 _ T H _ x t C _ m m _ t t a l P 1 _ T H _ x t C _ m m _ t t a l P
2 _ T H _ x t C _ m m _ t t a l P 2 _ T H _ x t C _ m m _ t t a l P
1 _ T W _ x t C _ m m _ t t a l P 1 _ T W _ x t C _ m m _ t t a l P
2 _ T W _ x t C _ m m _ t t a l P 2 _ T W _ x t C _ m m _ t t a l P
T H _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K T H _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K
d A _ n i a r b e l o h W _ m m _ a m a y a t a K d A _ n i a r b e l o h W _ m m _ a m a y a t a K
T W _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K T W _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K
T H _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K T H _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K
T W _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K T W _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K
G g I _ D 7 4 T _ g h _ z e v a h C s o l l a b e C G g I _ D 7 4 T _ g h _ z e v a h C s o l l a b e C − −
t p n I _ 1 _ 2 N R _ m m _ n e h S t p n I _ 1 _ 2 N R _ m m _ n e h S
t p n I _ 2 _ 2 N R _ m m _ n e h S t p n I _ 2 _ 2 N R _ m m _ n e h S
t p n I _ 2 _ C S N _ g h _ y e n t o C t p n I _ 2 _ C S N _ g h _ y e n t o C
t p n I _ 3 _ C S N _ g h _ y e n t o C t p n I _ 3 _ C S N _ g h _ y e n t o C
t p n I _ C S P i _ g h _ n a h t a g u S t p n I _ C S P i _ g h _ n a h t a g u S
t p n I _ T H _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K t p n I _ T H _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K
t p n I _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K t p n I _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K
t p n I _ T W _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K t p n I _ T W _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K
t p n I _ T H _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K t p n I _ T H _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K
t p n I _ T W _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K t p n I _ T W _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K GO:0012501 progr GO:0012501 iontr GO:0034220 regulation of small GO:0051056 GTP plasma membr GO:0005886 cell GO:0007155 adhesion cell GO:0034329 junctionassemb cytosk GO:0007010 e GO:0030198 axon guidance GO:0007411 projectionneuron de GO:0031175 cell GO:0016477 migr diff neuron GO:0030182 canonical GO:0060070 Wnt signalingpathw regulation of cell GO:0042127 prolif cell GO:0007049 cycle regulation of tr GO:0006355 e regulation of gene GO:0010468 organizationchromosome GO:0051276 processprotein metabolic GO:0019538 protein modification processGO:0036211 RNA GO:0008380 splicing regulation of RNA GO:0051252 metabolic process progr GO:0012501 iontr GO:0034220 regulation of small GO:0051056 GTP plasma membr GO:0005886 cell GO:0007155 adhesion cell GO:0034329 junctionassemb cytosk GO:0007010 e GO:0030198 axon guidance GO:0007411 projection neuron de GO:0031175 cell GO:0016477 migr diff neuron GO:0030182 canonical GO:0060070 Wnt signalingpathw regulation of cell GO:0042127 prolif cell GO:0007049 cycle regulation of tr GO:0006355 e regulation of gene GO:0010468 organizationchromosome GO:0051276 processprotein metabolic GO:0019538 protein modification processGO:0036211 RNA GO:0008380 splicing regulation of RNA GO:0051252 metabolic process xtr xtr acellular matr acellular matr ansmembr ansmembr ammed cell ammed death cell ammed death eleton organization eleton organization ation ation erentiation erentiation anscr anscr ane ane ane tr ane tr ix organization ix organization xpression xpression iption, DNA−templated iption, DNA−templated ly ly v v anspor anspor elopment elopment bioRxiv preprint doi: https://doi.org/10.1101/375238; this versioner posted July 23, 2018. The copyright holderer for this preprint (which was not ase mediated signal ase mediated tr signal ase mediated tr certified by peer review) is the author/funder, who has granted bioRxivation a license to display the preprint in perpetuity.ation It is made available under t t
aCC-BYa 4.0 International license. a y y Wade et al. CHD8 Regulates Cellular Function Genes ansduction ansduction
FigureFigure 5 5 0 2 4 6 8 10 0 2 4 6 8 10
0 2 4 6 8 10 0 2 4 6 8 10
D 7 4 T _ g h _ z e v a h C s o l l a b e C D 7 4 T _ g h _ z e v a h C s o l l a b e C
−
− 1 _ 2 N R _ m m _ n e h S 1 _ 2 N R _ m m _ n e h S
2 _ 2 N R _ m m _ n e h S 2 _ 2 N R _ m m _ n e h S
1 _ C S E _ m m _ t l u e v e l u e i D e d 1 _ C S E _ m m _ t l u e v e l u e i D e
d
ansduction ansduction 2 _ C S E _ m m _ t l u e v e l u e i D e d 2 _ C S E _ m m _ t l u e v e l u e i D e d
2 _ C S N _ g h _ y e n t o C 2 _ C S N _ g h _ y e n t o C
3 _ C S N _ g h _ y e n t o C 3 _ C S N _ g h _ y e n t o
C Immunoprecipitated Immunoprecipitated
l y h t e B _ C S P i _ g h _ n a h t a g u S l y h t e B _ C S P i _ g h _ n a h t a g u S
7 1 4 0 6 s u v o N _ C S P i _ g h _ n a h t a g u S 7 1 4 0 6 s u v o N _ C S P i _ g h _ n a h t a g u
S 8 1 4 0 6 s u v o N _ C S P i _ g h _ n a h t a g u S 8 1 4 0 6 s u v o N _ C S P i _ g h _ n a h t a g u S
y y
1 _ 7 1 E _ _ x t C _ m m _ y e n t o C 1 _ 7 1 E _ _ x t C _ m m _ y e n t o a C a
t t
2 _ 7 1 E _ x t C _ m m _ y e n t o C 2 _ 7 1 E _ x t C _ m m _ y e n t o
C by EnrichedFDR
n i a r b e r o F _ m m _ s r e p m o G n i a r b e r o F _ m m _ s r e p m o G
ation ation
ase mediatedtr signal ase mediatedtr signal 1 _ T H _ x t C _ m m _ t t a l P 1 _ T H _ x t C _ m m _ t t a l P
er er elopment elopment
anspor anspor
2 _ T H _ x t C _ m m _ t t a l P 2 _ T H _ x t C _ m m _ t t a l v P v
ly ly iption, DNA−templated iption, DNA−templated iption, xpression xpression
1 _ T W _ x t C _ m m _ t t a l P 1 _ T W _ x t C _ m m _ t t a l ix organization ix P organization ix
ane tr ane tr 2 _ T W _ x t C _ m m _ t t a l P 2 _ T W _ x t C _ m m _ t t a l ane P ane
anscr anscr T H _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K T H _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K
d A _ n i a r b e l o h W _ m m _ a m a y a t a K d A _ n i a r b e l o h W _ m m _ a m a y a t a erentiation K erentiation
ation ation T W _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K T W _ d A _ n i a r b e l o h W _ m m _ a m a y a t a
K
eleton organization eleton organization
T H _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K T H _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a ammed death cell K ammed death cell
ansmembr ansmembr
acellular matr acellular matr acellular T W _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K T W _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K
xtr xtr
G g I _ D 7 4 T _ g h _ z e v a h C s o l l a b e C G g I _ D 7 4 T _ g h _ z e v a h C s o l l a b e C − −
t p n I _ 1 _ 2 N R _ m m _ n e h S t p n I _ 1 _ 2 N R _ m m _ n e h S
t p n I _ 2 _ 2 N R _ m m _ n e h S t p n I _ 2 _ 2 N R _ m m _ n e h
S
t p n I _ 2 _ C S N _ g h _ y e n t o C t p n I _ 2 _ C S N _ g h _ y e n t o C
Control Control
t p n I _ 3 _ C S N _ g h _ y e n t o C t p n I _ 3 _ C S N _ g h _ y e n t o C
t p n I _ C S P i _ g h _ n a h t a g u S t p n I _ C S P i _ g h _ n a h t a g u
S
GO:0012501 progr GO:0034220 tr ion GO:0051056 GTP regulation of small GO:0005886 plasma membr adhesion GO:0007155 cell assemb junction GO:0034329 cell GO:0007010 cytosk GO:0030198 e GO:0007411 axon guidance GO:0031175 neuron de projection migr GO:0016477 cell GO:0030182 neuron diff pathw signaling GO:0060070 Wnt canonical prolif GO:0042127 regulation of cell cycle GO:0007049 cell GO:0006355 regulation of tr GO:0010468 regulation of gene e GO:0051276 chromosome organization GO:0019538 protein metabolic process GO:0036211 process protein modification splicing GO:0008380 RNA metabolic process GO:0051252 regulation of RNA GO:0012501 progr GO:0034220 tr ion GO:0051056 GTP regulation of small GO:0005886 plasma membr adhesion GO:0007155 cell assemb junction GO:0034329 cell GO:0007010 cytosk GO:0030198 e GO:0007411 axon guidance GO:0031175 neuron de projection migr GO:0016477 cell GO:0030182 neuron diff pathw signaling GO:0060070 Wnt canonical prolif GO:0042127 regulation of cell cycle GO:0007049 cell GO:0006355 regulation of tr GO:0010468 regulation of gene e GO:0051276 chromosome organization GO:0019538 protein metabolic process GO:0036211 process protein modification splicing GO:0008380 RNA metabolic process GO:0051252 regulation of RNA t p n I _ T H _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K t p n I _ T H _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K
t p n I _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K t p n I _ d A _ n i a r b e l o h W _ m m _ a m a y a t a t p n I _ T W _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K t p n I _ T W _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K K
t p n I _ T W _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K t p n I _ T W _ d A _ n i a r b e l o h W _ m m _ a m a y a t a t p n I _ T H _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K t p n I _ T H _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K K
t p n I _ T H _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K t p n I _ T H _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a t p n I _ T W _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K t p n I _ T W _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K K
D 7 4 T _ g h _ z e v a h C s o l l a b e C − t p n I _ T W _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K t p n I _ T W _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a t p n I _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K t p n I _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K K
D 7 4 T _ g h _ z e v a h C s o l l a b e C
− 1 _ 2 N R _ m m _ n e h
t p n I _ T H _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K t p n I _ T H _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K S
GO:0012501 progr GO:0012501 iontr GO:0034220 regulation of small GO:0051056 GTP plasma membr GO:0005886 cell GO:0007155 adhesion cell GO:0034329 junctionassemb cytosk GO:0007010 e GO:0030198 axon guidance GO:0007411 projection neuron de GO:0031175 cell GO:0016477 migr diff neuron GO:0030182 canonical GO:0060070 Wnt signalingpathw regulation of cell GO:0042127 prolif cell GO:0007049 cycle regulation of tr GO:0006355 e regulation of gene GO:0010468 organizationchromosome GO:0051276 processprotein metabolic GO:0019538 protein modification processGO:0036211 RNA GO:0008380 splicing regulation of RNA GO:0051252 metabolic process progr GO:0012501 iontr GO:0034220 regulation of small GO:0051056 GTP plasma membr GO:0005886 cell GO:0007155 adhesion cell GO:0034329 junctionassemb cytosk GO:0007010 e GO:0030198 axon guidance GO:0007411 projection neuron de GO:0031175 cell GO:0016477 migr diff neuron GO:0030182 canonical GO:0060070 Wnt signalingpathw regulation of cell GO:0042127 prolif cell GO:0007049 cycle regulation of tr GO:0006355 e regulation of gene GO:0010468 organizationchromosome GO:0051276 processprotein metabolic GO:0019538 protein modification processGO:0036211 RNA GO:0008380 splicing regulation of RNA GO:0051252 metabolic process
1 _ 2 N R _ m m _ n e h S
2 _ 2 N R _ m m _ n e h t p n I _ C S P i _ g h _ n a h t a g u S t p n I _ C S P i _ g h _ n a h t a g u S S
2 _ 2 N R _ m m _ n e h S
1 _ C S E _ m m _ t l u e v e l u e i D e
t p n I _ 3 _ C S N _ g h _ y e n t o C All Peaks t p n I _ 3 _ C S N _ g h _ y e n t o C d Top 2000 Peaks
1 _ C S E _ m m _ t l u e v e l u e i D e
d
2 _ C S E _ m m _ t l u e v e l u e i D e t p n I _ 2 _ C S N _ g h _ y e n t o C t p n I _ 2 _ C S N _ g h _ y e n t o C d
D 7 4 T _ g h _ z e v a h C s o l l a b e C D 7 4 T _ g h _ z e v a h C s o l l a b e C 2 _ C S E _ m m _ t l u e v e l u e i D e d
− −
2 _ C S N _ g h _ y e n t o t p n I _ 2 _ 2 N R _ m m _ n e h S t p n I _ 2 _ 2 N R _ m m _ n e h S C
1 _ 2 N R _ m m _ n e h S 1 _ 2 N R _ m m _ n e h S 2 _ C S N _ g h _ y e n t o C
3 _ C S N _ g h _ y e n t o
t p n I _ 1 _ 2 N R _ m m _ n e h S t p n I _ 1 _ 2 N R _ m m _ n e h S C
2 _ 2 N R _ m m _ n e h S 2 _ 2 N R _ m m _ n e h S 3 _ C S N _ g h _ y e n t o C
l y h t e B _ C S P i _ g h _ n a h t a g u
G g I _ D 7 4 T _ g h _ z e v a h C − s o l l a b e C G g I _ D 7 4 T _ g h _ z e v a h C − s o l l a b e C S
xtr xtr
1 _ C S E _ m m _ t l u e v e l u e i D e d 1 _ C S E _ m m _ t l u e v e l u e i D e d l y h t e B _ C S P i _ g h _ n a h t a g u S
7 1 4 0 6 s u v o N _ C S P i _ g h _ n a h t a g u
T W _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K T W _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K S
acellular matr acellular matr
2 _ C S E _ m m _ t l u e v e l u e i D e d 2 _ C S E _ m m _ t l u e v e l u e i D e d 7 1 4 0 6 s u v o N _ C S P i _ g h _ n a h t a g u
S ansmembr ansmembr
cell ammed death cell ammed death 8 1 4 0 6 s u v o N _ C S P i _ g h _ n a h t a g u
T H _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K T H _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K S
eleton organization eleton organization
2 _ C S N _ g h _ y e n t o C 2 _ C S N _ g h _ y e n t o C 8 1 4 0 6 s u v o N _ C S P i _ g h _ n a h t a g u S
1 _ 7 1 E _ _ x t C _ m m _ y e n t o
T W _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K ation T W _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K C ation
3 _ C S N _ g h _ y e n t o C 3 _ C S N _ g h _ y e n t o C 1 _ 7 1 E _ _ x t C _ m m _ y e n t o
C Immunoprecipitated Immunoprecipitated
2 _ 7 1 E _ x t C _ m m _ y e n t o
d A _ n i a r b e l o h W _ m m _ a m a y a t a K erentiation d A _ n i a r b e l o h W _ m m _ a m a y a t a K C erentiation
l y h t e B _ C S P i _ g h _ n a h t a g u S l y h t e B _ C S P i _ g h _ n a h t a g u S 2 _ 7 1 E _ x t C _ m m _ y e n t o
C
n i a r b e r o F _ m m _ s r e p m o
T H _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K T H _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K G
7 1 4 0 6 s u v o N _ C S P i _ g h _ n a h t a g u S 7 1 4 0 6 s u v o N _ C S P i _ g h _ n a h t a g u S
n i a r b e r o F _ m m _ s r e p m o
G anscr anscr
ane ane
1 _ T H _ x t C _ m m _ t t a l
2 _ T W _ x t C _ m m _ t t a l P ane tr 2 _ T W _ x t C _ m m _ t t a l P P ane tr Enriched by EnrichedProportion
8 1 4 0 6 s u v o N _ C S P i _ g h _ n a h t a g u S 8 1 4 0 6 s u v o N _ C S P i _ g h _ n a h t a g u S 1 _ T H _ x t C _ m m _ t t a l
P ix organization ix organization
2 _ T H _ x t C _ m m _ t t a l
1 _ T W _ x t C _ m m _ t t a l P 1 _ T W _ x t C _ m m _ t t a l P P
1 _ 7 1 E _ _ x t C _ m m _ y e n t o C 1 _ 7 1 E _ _ x t C _ m m _ y e n t o C 2 _ T H _ x t C _ m m _ t t a l
P xpression xpression iption, DNA−templated iption, DNA−templated
ly ly
1 _ T W _ x t C _ m m _ t t a l
2 _ T H _ x t C _ m m _ t t a l P v 2 _ T H _ x t C _ m m _ t t a l P P v
anspor anspor
2 _ 7 1 E _ x t C _ m m _ y e n t o elopment C elopment 2 _ 7 1 E _ x t C _ m m _ y e n t o C 1 _ T W _ x t C _ m m _ t t a l
P er er
2 _ T W _ x t C _ m m _ t t a l
1 _ T H _ x t C _ m m _ t t a l P signal ase mediated tr 1 _ T H _ x t C _ m m _ t t a l P P signal ase mediated tr
ation ation
n i a r b e r o F _ m m _ s r e p m o G n i a r b e r o F _ m m _ s r e p m o G 2 _ T W _ x t C _ m m _ t t a l
P T H _ d A _ n i a r b e l o h W _ m m _ a m a y a t a
n i a r b e r o F _ m m _ s r e p m o G n i a r b e r o F _ m m _ s r e p m o G K
1 _ T H _ x t C _ m m _ t t a l P 1 _ T H _ x t C _ m m _ t t a l P T H _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K
d A _ n i a r b e l o h W _ m m _ a m a y a t a
2 _ 7 1 E _ x t C _ m m _ y e n t o C t 2 _ 7 1 E _ x t C _ m m _ y e n t o C K t
2 _ T H _ x t C _ m m _ t t a l P 2 _ T H _ x t C _ m m _ t t a l P d A _ n i a r b e l o h W _ m m _ a m a y a t a K
a a
T W _ d A _ n i a r b e l o h W _ m m _ a m a y a t a
1 _ 7 1 E _ _ x t C _ m m _ y e n t o C 1 _ 7 1 E _ _ x t C _ m m _ y e n t o C K
y y 1 _ T W _ x t C _ m m _ t t a l P 1 _ T W _ x t C _ m m _ t t a l P T W _ d A _ n i a r b e l o h W _ m m _ a m a y a t a
K
T H _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a
8 1 4 0 6 s u v o N _ C S P i _ g h _ n a h t a g u S 8 1 4 0 6 s u v o N _ C S P i _ g h _ n a h t a g u S K
2 _ T W _ x t C _ m m _ t t a l P 2 _ T W _ x t C _ m m _ t t a l P T H _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K
T W _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a
7 1 4 0 6 s u v o N _ C S P i _ g h _ n a h t a g u S 7 1 4 0 6 s u v o N _ C S P i _ g h _ n a h t a g u S K
T H _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K T H _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K T W _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K
G g I _ D 7 4 T _ g h _ z e v a h C s o l l a b e C
l y h t e B _ C S P i _ g h _ n a h t a g u S l y h t e B _ C S P i _ g h _ n a − h t a g u S
d A _ n i a r b e l o h W _ m m _ a m a y a t a K d A _ n i a r b e l o h W _ m m _ a m a y a t a K G g I _ D 7 4 T _ g h _ z e v a h C s o l l a b e C −
t p n I _ 1 _ 2 N R _ m m _ n e h
3 _ C S N _ g h _ y e n t o C 3 _ C S N _ g h _ y e n t o C S
T W _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K T W _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K t p n I _ 1 _ 2 N R _ m m _ n e h S
t p n I _ 2 _ 2 N R _ m m _ n e h
2 _ C S N _ g h _ y e n t o C 2 _ C S N _ g h _ y e n t o C S
T H _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K T H _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K t p n I _ 2 _ 2 N R _ m m _ n e h
S t p n I _ 2 _ C S N _ g h _ y e n t o
2 _ C S E _ m m _ t l u e v e l u e i D e d ansduction 2 _ C S E _ m m _ t l u e v e l u e i D e d C ansduction
T W _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K T W _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K t p n I _ 2 _ C S N _ g h _ y e n t o C
t p n I _ 3 _ C S N _ g h _ y e n t o
1 _ C S E _ m m _ t l u e v e l u e i D e d 1 _ C S E _ m m _ t l u e v e l u e i D e d C
G g I _ D 7 4 T _ g h _ z e v a h C s o l l a b e C − G g I _ D 7 4 T _ g h _ z e v a h C s o l l a b e C − t p n I _ 3 _ C S N _ g h _ y e n t o C
t p n I _ C S P i _ g h _ n a h t a g u
2 _ 2 N R _ m m _ n e h S 2 _ 2 N R _ m m _ n e h S S
t p n I _ 1 _ 2 N R _ m m _ n e h S t p n I _ 1 _ 2 N R _ m m _ n e h S t p n I _ C S P i _ g h _ n a h t a g u
S
t p n I _ T H _ d A _ n i a r b e l o h W _ m m _ a m a y a t a
1 _ 2 N R _ m m _ n e h S 1 _ 2 N R _ m m _ n e h S K
t p n I _ 2 _ 2 N R _ m m _ n e h S t p n I _ 2 _ 2 N R _ m m _ n e h S t p n I _ T H _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K
t p n I _ d A _ n i a r b e l o h W _ m m _ a m a y a t a
D 7 4 T _ g h _ z e v a h C − s o l l a b e C D 7 4 T _ g h _ z e v a h C − s o l l a b e C K t p n I _ 2 _ C S N _ g h _ y e n t o C t p n I _ 2 _ C S N _ g h _ y e n t o C t p n I _ d A _ n i a r b e l o h W _ m m _ a m a y a t a
K Control Control t p n I _ T W _ d A _ n i a r b e l o h W _ m m _ a m a y a t a
K
t p n I _ 3 _ C S N _ g h _ y e n t o C t p n I _ 3 _ C S N _ g h _ y e n t o C t p n I _ T W _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K
t p n I _ T H _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a
0 2 4 6 8 10 K 0 2 4 6 8 10
t p n I _ C S P i _ g h _ n a h t a g u S t p n I _ C S P i _ g h _ n a h t a g u S t p n I _ T H _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K
t p n I _ T W _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K
t p n I _ T H _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K t p n I _ T H _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K t p n I _ T W _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a
K
GO:0012501 progr GO:0012501 iontr GO:0034220 regulation of small GO:0051056 GTP plasma membr GO:0005886 cell GO:0007155 adhesion cell GO:0034329 junctionassemb cytosk GO:0007010 e GO:0030198 axon guidance GO:0007411 projection neuron de GO:0031175 cell GO:0016477 migr diff neuron GO:0030182 canonical GO:0060070 Wnt signalingpathw regulation of cell GO:0042127 prolif cell GO:0007049 cycle regulation of tr GO:0006355 e regulation of gene GO:0010468 organizationchromosome GO:0051276 processprotein metabolic GO:0019538 protein modification processGO:0036211 RNA GO:0008380 splicing regulation of RNA GO:0051252 metabolic process t p n I _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K t p n I _ d A _ n i a r b e l o h W _ m m _ a m a y a t a
GO:0012501 progr GO:0012501 K iontr GO:0034220 regulation of small GO:0051056 GTP plasma membr GO:0005886 cell GO:0007155 adhesion cell GO:0034329 junctionassemb cytosk GO:0007010 e GO:0030198 axon guidance GO:0007411 projection neuron de GO:0031175 cell GO:0016477 migr diff neuron GO:0030182 canonical GO:0060070 Wnt signalingpathw regulation of cell GO:0042127 prolif cell GO:0007049 cycle regulation of tr GO:0006355 e regulation of gene GO:0010468 organizationchromosome GO:0051276 processprotein metabolic GO:0019538 protein modification processGO:0036211 RNA GO:0008380 splicing regulation of RNA GO:0051252 metabolic process
t p n I _ T W _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K t p n I _ T W _ d A _ n i a r b e l o h W _ m m _ a m a y a t a K
t p n I _ T H _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K t p n I _ T H _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K
t p n I _ T W _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K t p n I _ T W _ 4 1 E _ n i a r b e l o h W _ m m _ a m a y a t a K GO:0012501 progr GO:0012501 iontr GO:0034220 regulation of small GO:0051056 GTP plasma membr GO:0005886 cell GO:0007155 adhesion cell GO:0034329 junctionassemb cytosk GO:0007010 e GO:0030198 axon guidance GO:0007411 projection neuron de GO:0031175 cell GO:0016477 migr diff neuron GO:0030182 canonical GO:0060070 Wnt signalingpathw regulation of cell GO:0042127 prolif cell GO:0007049 cycle regulation of tr GO:0006355 e regulation of gene GO:0010468 organizationchromosome GO:0051276 processprotein metabolic GO:0019538 protein modification processGO:0036211 RNA GO:0008380 splicing regulation of RNA GO:0051252 metabolic process progr GO:0012501 iontr GO:0034220 regulation of small GO:0051056 GTP plasma membr GO:0005886 cell GO:0007155 adhesion cell GO:0034329 junctionassemb cytosk GO:0007010 e GO:0030198 axon guidance GO:0007411 projection neuron de GO:0031175 cell GO:0016477 migr diff neuron GO:0030182 canonical GO:0060070 Wnt signalingpathw regulation of cell GO:0042127 prolif cell GO:0007049 cycle regulation of tr GO:0006355 e regulation of gene GO:0010468 organizationchromosome GO:0051276 processprotein metabolic GO:0019538 protein modification processGO:0036211 RNA GO:0008380 splicing regulation of RNA GO:0051252 metabolic process xtr xtr acellular matr acellular matr ansmembr
cell ammed death ansmembr ammed cell ammed death eleton organization eleton organization ation ation erentiation erentiation
xtr xtr acellular matr anscr acellular matr ane anscr ansmembr ane ane tr ammed cell ammed death ansmembr ammed cell ammed death ane tr eleton organization eleton organization ix organization
ix organization ation ation xpression iption, DNA−templated ly xpression v iption, DNA−templated ly erentiation v erentiation anspor elopment er anspor elopment er ase mediated signal ase mediated tr ation ase mediated signal ase mediated tr ation
anscr anscr ane ane ane tr ane tr ix organization ix organization t t xpression iption, DNA−templated a ly xpression v iption, DNA−templated a ly
v y anspor elopment y er anspor elopment er ase mediated signal ase mediated tr ation ase mediated signal ase mediated tr ation t t a a y y ansduction ansduction ansduction ansduction 0.2 0.4 0.6 0.8 0.05 0.1 0.15 0.2
0.2 0.4 0.6 0.8 0.05 0.1 0.15 0.2 Unexplained specificity of CHD8 binding near gene regulator promoters. (Left) Enrichment of terms according to an FDR < 0.05 cutoff. (Right) Enrichment of terms according to proportion ranging from 0, not enriched, to 1, all genes in the term category. (Top) Analysis of the top 2000 significant peaks. (Bottom) Analysis of all peaks meeting a p < 0.00001 MACS2 significance level.
24 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes Figure 6 Figure 6
GompersGomper DE xs_Full Platt ChIP GompersGomper DE x Gomperss_Full ChIP GompersGomper DE x Katayamas_Full ChIP
4 4 4 . . . 0 0 0
2 2 2 . . .
0 0 0
C C C F 0 F F 0 0 . . . g g g
0 0 0 o o o L L L 2 2 2 . . . 0 0 0 − − − 4 4 4 . . . 0 0 0 − − − Gompers_Full Gompers_Full Gompers_Full
5 5 5 1 1 1
0 0 0 1 1 1
5 M 5 5 M M K K K P P P R R R g g g o o o 0 0 0 L L L
5 5 5 − − −
0 0 0 1 1 1 − − −
Expressed Genes Rank 1001-2000 Bound Top 250 Bound
All Bound Genes Rank 251-1000 Bound
CHD8 regulates differentially expressed genes with high-confidence CHD8 binding. (Left) Comparison between the full Gompers et al. 2017 Chd8 heterozygous mouse model differential expression gene set (DGE) and the Platt et al. 2017 Chd8 ChIP-seq dataset. (Middle) Comparison between the Gompers et al. DGE and Gompers et al. Chd8 ChIP-seq dataset. (Right) Comparison between the Gompers et al. DGE and Katayama et al. 2016 Chd8 ChIP-seq dataset. (Top) Change in expression of genes according to CHD8 binding compared to wild-type littermates. (Bottom) Change in sequence coverage of genes according to CHD8 binding. Boxes were plotted according to CHD8 binding affinity bins: all genes meeting a 1 count per million sequencing coverage threshold included in DEG analysis (Expressed Genes), any genes having CHD8 binding (All Bound Genes), and all genes having binding ranked according to CHD8 peak significance (Top 250 Bound, Rank 251-1000 Bound, Rank 1001-2000 Bound).
25 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes Supplementary Figure 1 Supplementary Figure 1
A D
B E
C F
Analyzing individual CHD8 model gene expression and pairwise comparisons through the Shiny interactive web browser. (A) The CHD8 RNA-seq Shiny app can generate a principle component analysis (PCA) scatter plot with any of the RNA-seq datasets loaded onto the app. You can tailor the plot according to several parameters including sample number, principle component, gene of interest, dataset, timepoint, sex, model organism, mouse or cell line, and genotype. This plot
26 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes
shows the Cotney et al. 2015 and Durak et al. 2016 RNA-seq datasets plotted according to the experimental design, which is in this case the control and experimental constructs. A multidimensional scaling (MDS) plot (B) and log fold change differential gene expression bar plot (C) generated using Shiny is also shown with experimental design chosen as the display parameters. Almost all of the same display parameters for the PCA plot are available for the MDS and box plots. Supplementary tables showing metadata (D) and gene counts (E) are also available through the Shiny app. (F) Table showing log fold gene expression changes and significance values for individual genes between the Cotney et al. 2015 and Durak et al. 2016 RNA-seq datasets. Heat maps and scatter plots of gene expression changes are also available. All plots and tables generated using Shiny can be downloaded from the app. Datasets can be analyzed using pseudo counts or relative expression.
Supplementary Figure 2
(See Supplementary Figure File)
Full list of terms from the gene ontology analysis using goseq. Terms were selected from this list to create Figure 2. All RNA-seq datasets were included in this figure. All terms included met an FDR < 0.05 cutoff.
27 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes
SupplementarySupplementary Figure 3Figure 3
5 Up GO:0051252 regulation of RNA metabolic process Up GO:0008380 RNA splicing Up GO:0036211 protein modification process 4 Up GO:0019538 protein metabolic process Up GO:0051276 chromosome organization 3 Up GO:0010468 regulation of gene expression Up GO:0006355 regulation of transcription, DNA−templated 2 Up GO:0007049 cell cycle Up GO:0042127 regulation of cell proliferation 1 Up GO:0060070 canonical Wnt signaling pathway Up GO:0030182 neuron differentiation Up GO:0016477 cell migration 0 Up GO:0031175 neuron projection development Up GO:0007411 axon guidance Up GO:0030198 extracellular matrix organization Up GO:0007010 cytoskeleton organization Up GO:0034329 cell junction assembly Up GO:0007155 cell adhesion Up GO:0005886 plasma membrane Up GO:0051056 regulation of small GTPase mediated signal transduction Up GO:0034220 ion transmembrane transport Up GO:0012501 programmed cell death Down GO:0051252 regulation of RNA metabolic process Down GO:0008380 RNA splicing Down GO:0036211 protein modification process Down GO:0019538 protein metabolic process Down GO:0051276 chromosome organization Down GO:0010468 regulation of gene expression Down GO:0006355 regulation of transcription, DNA−templated Down GO:0007049 cell cycle Down GO:0042127 regulation of cell proliferation Down GO:0060070 canonical Wnt signaling pathway Down GO:0030182 neuron differentiation Down GO:0016477 cell migration Down GO:0031175 neuron projection development Down GO:0007411 axon guidance Down GO:0030198 extracellular matrix organization Down GO:0007010 cytoskeleton organization Down GO:0034329 cell junction assembly Down GO:0007155 cell adhesion Down GO:0005886 plasma membrane Down GO:0051056 regulation of small GTPase mediated signal transduction Down GO:0034220 ion transmembrane transport Down GO:0012501 programmed cell death D W W D S G C K G G S G W K P P G K K K W P S C P P P h a u a l l a a a l h l l l u u o o o o o o o a a a a a a a a a i l e t g t t t t e r r t t m m m m m k t t t t t t n n n a a a a a n n a a t t t t t t n a n i ______g g g y y y y y e n e k k p p p p p _ t _
a a a a a S H A H V N _ _ _ h y y _ _ s e e e e e R R m m m m m N O N _ _ o a s s t m T r r r r i r y u r p N s s s s s N s s h h A n n p c P e i a a a a a r y p a _ _ _ _ _ h h g C C _ _ 2 2 _ A _ u _ _ _ _ _ C g
_ t P E P E E C C a F s _ _ P u P r H H c E P E E E _ P _ o i 0 1 5 1 1 n s u s m H H c C v v P 1 9 1 1 1 D D P v n _ 7 6 2 4 o g h l _ a a D D l 8 1 6 2 0 H _ a v _ 8 8 P . _ _ . . v i C C P l l d
5 5 5 _ _ _ _ _ a u P u l 8 8 a P . . P P D u v H H v _ 2 1 _ _ _ P P P P P l e e . . l v u a v e v v a u 8 C G P _ _ P P P D D a v v v v v a e l a a l e _ u P P v _ u a a a a a v v v _ l l 8 8 l l
u P u e u u a a a a P e l l l l l v v P _ _ u u u u u e e e e l v a a l l l v u u u u v P P e e e e e a l l a a e u u e e e v v l l u l e e
a a u u e l l e e u u e e
Enrichment of gene regulation and neurodevelopmental ontology terms meeting a goseq p < 0.05 significance level in all Up- and Down-regulated datasets. (Top) Upregulated gene ontology enrichment. (Bottom) Downregulated gene ontology enrichment.
28 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes
SupplementarySupplement Figureary Figure 4 4
ELF1 ELK1 E2F CTCF YY1
Ceballos ✓ ✓ ✓ Cotney Hum ✓ ✓ Sug. 60417 ✓
Sug. 60418 ✓ ✓
Sug. Bethyl ✓ Gompers ✓
Platt HT ✓ ✓ ✓
Platt WT ✓ ✓
Katayama Ad ✓ ✓
Katayama E14 ✓ ✓
De Dieu ✓ ✓
Cotney Mouse ✓ ✓
Shen ✓ ✓
No obvious primary motif associated with CHD8 binding. Each dataset was analyzed using HOMER to look for common motifs enriched in CHD8 ChIP-seq datasets. ELF1, ELK1, E2F, CTCF, and YY1 transcription factors were the motifs that were commonly represented across datasets. Ceballos – Ceballos-Chavez et al. 2015 ChIP-seq dataset, Sug. 60417, 60418, Bethyl – Sugathan et al. 2014 ChIP-seq datasets split according to antibody used (60417, NB100-60417; 60418, NB100-60418), Gompers – Gompers et al. 2017 ChIP-seq dataset, Platt HT, WT – Platt et al. 2017 ChIP-seq datasets split according to genotype of the samples (HT, heterozygous; WT, wild-type), Katayama – Katayama et al. 2016 ChIP-seq datasets split according to age of the samples (Ad, postnatal day 91; E14, embryonic day 14), De Dieu – De Dieulevuelt et al. 2016 ChIP-seq dataset, Cotney Hum, Mouse – Cotney et al. 2015 ChIP-seq datasets split according to model organism (Hum, human), Shen – Shen et al. 2015 ChIP-seq dataset.
29 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes Supplementary Figure 5 Supplementary Figure 5
Wang_Neuron Wang_NPC Wang_Organoid Sugathan_Full Cotney_Full
5 5 5 5 5 1 1 1 1 1 0 0 0 0 0
1 1 1 1 1
5 5 5 5 5 M M M M M
K K K K K P P P P P R R R R R
g g g g g o o o o o 0 0 0 0 0 L L L L L
5 5 5 5 5 − − − − − 0 0 0 0 0
1 1 1 1 1 − − − − −
Shen_RN2_shCHD8 Shen_RN2_sgCHD8 Wilkinson_siCHD8 Katayama_Full Platt_Full
5 5 5 5 5 1 1 1 1 1
0 0 0 0 0 1 1 1 1 1
5 5 5 5 5 M M M M M K K K K K P P P P P R R R R R g g g g g o o o o o 0 0 0 0 0 L L L L L 5 5 5 5 5 − − − −
−
0 0 0 0 0 1 1 1 1 1 − − − − −
Expressed Genes Rank 1001-2000 Bound Top 250 Bound
All Bound Genes Rank 251-1000 Bound
Chd8 regulates highly expressed genes. Each dataset is labelled showing changes in sequencing coverage according to changes in CHD8 binding affinity. The Chd8 ChIP-seq dataset used was from Platt et al. 2017. Full models for each dataset were chosen as they exhibited similar signal as the individual timepoint or brain region datasets.
30 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes
Supplementary Figure 6
Remaining fold change plots from the CHD8 binding by differential gene expression comparison analysis. All datasets were analyzed using the Platt et al. 2017 Chd8 ChIP-seq dataset. Datasets are loosely organized based on overlap between downregulated genes, no clear trend, or upregulated genes from top to bottom, which sometimes spanned multiple rows.
31 bioRxiv preprint doi: https://doi.org/10.1101/375238; this version posted July 23, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY 4.0 International license. Wade et al. CHD8 Regulates Cellular Function Genes
SupplementarySupplementary Figure 7 Figure 7
Durak_Full Durak_Full Durak_Full Durak DE x Platt ChIP Durak DE x Gompers ChIP Durak DE x Katayama ChIP
4 4 4 . . . 0 0 0
2 2 2 . . . 0 0 0 C C C F F F 0 0 0 . . . g g g 0 0 0 o o o L L L 2 2 2 . . . 0 0 0 − − −
4 4 4 . . . 0 0 0 − − − Durak_Full Durak_Full Durak_Full
5 5 5 1 1 1
0
0 0 1 1 1
5 M 5 5 M M K K K P P P R R R g g g o o o 0 0 0 L L L
5 5 5 − − − 0 0 0 1 1 1 − − −
Expressed Genes Rank 1001-2000 Bound Top 250 Bound
All Bound Genes Rank 251-1000 Bound
Chd8 regulates genes differentially expressed with Chd8 knockdown. (Left) Comparison between the full Durak et al. 2016 Chd8 knockdown mouse model differential expression gene set (DGE) and the Platt et al. 2017 Chd8 ChIP-seq dataset. (Middle) Comparison between the Durak et al. DGE and Gompers et al. Chd8 ChIP-seq dataset. (Right) Comparison between the Durak et al. DGE and Katayama et al. 2016 Chd8 ChIP-seq dataset. (Top) Change in expression of all genes compared to wild-type littermates according to changes in CHD8 binding affinity. (Bottom) Changes in sequencing coverage of genes according to changes in CHD8 binding affinity.
32