Supplementary Information for:

Modeling genomic diversity and tumor dependency in malignant melanoma

Updated information pertaining to microarray data processing may be found at the following url: http://www.broad.mit.edu/melanoma

Supplementary Methods

Cultured melanoma cells

The melanoma short-term cultures (STCs) employed in this study were all derived from metastatic disease. In most cases, their melanocytic origins were confirmed by histological or molecular studies.

High-density SNP array hybridization and data processing

Genomic DNA from all samples was tracked with ABgene 2D barcode rack and single tube readers (ABgene). The well designation for samples was randomized, and in each plate there were 6 control wells that contained normal diploid DNA. After the genomic DNA was cut by StyI restriction enzyme (New England Biolabs), the DNA was ligated to an adaptor with a T4 ligase (New England Biolabs). PCR amplification was then done using an Applied Biosystems 9700 Thermal Cycler I and Titanium Taq (Clontech) to obtain 200-110 bp fragments. After the amplified DNA was pooled and concentrated, it was fragmented using DNaseI (Affymetrix, Inc.). The DNA was then end-labeled, denatured and hybridized to the probe sets in each well of the plate and washed. DNA was then hybridized to the microarrays according to the manufacturer’s directions. 50K XbaI SNP hybridization was performed as described previously.

Using the SNPFileCreator GenePattern module, a .snp file was created by normalizing intensities and modeling using PM-MM difference modeling method (Li, 2001). Each .snp file contained normalized raw intensities for samples along with SNP marker information and chromosomal position (based upon Affymetrix annotations and hg17 build of the sequence from the University of California, Santa Cruz: http://genome.ucsc.edu/).

Next, a series of manipulations were implemented to increase the signal–to-noise ratio. First, melanoma samples and normal samples from each of the .snp files were selected, sorted by chromosomal position, and merged into a single matrix after median adjusted batch-correction. In order to convert raw intensity data into copy-number values, normal samples with the highest similarity to a given tumor sample were chosen by median log2- transformed Euclidean distance across the corresponding markers. To remove normal samples with sub-optimal data quality, a Gaussian histogram analysis was generated so that copy number distribution could be examined. Here, high-quality normal samples typically show a single narrow peak centered at two copies, whereas hyperploid tumor samples typically have multiple peaks with a broader signal range (this method removes

1 samples with significant stromal admixture but may also exclude pure tumors that lack copy number abnormalities). The copy number data for each tumor sample was then re- computed, referencing the high-quality normal samples. Heatmap visualization of SNP array data uses a beta version of the GenePattern SNP visualizer (A full-version is expected soon, http://www.broad.mit.edu/cancer/software/genepattern/index.html).

Next, the GLAD segmentation algorithm (1) was employed to smooth the copy number data with default parameters. Noisy samples reflecting technical artifacts were removed at this step using a distribution metric of the segmented data. SNP markers in regions of documented copy-number variants (http://projects.tcag.ca/variation/, October 2006) were removed from subsequent analysis, thereby reducing the density of markers. To avoid bias from using multiple clones from a single patient, such samples (where present) were identified through a genotype comparison. Next, samples with the least noise from among the clones were chosen based on noise between neighboring probes. In combining 250K and 50K data sets, both data sets were aligned by genomic position, and missing cells in the matrix were filled in with the mean of neighboring values. All source code was written in MATLAB R2006b. Further details of processing can be found in a forthcoming paper by Beroukhim et al. (manuscript submitted).

RNA preparation and expression profiling RNA was extracted from melanoma short term cultures grown to 80% confluency with 1.0 mL of Trizol Reagent with 200 µL of chloroform. After shaking and incubation at 30°C for 10 minutes, the samples were centrifuged at 13,000 RPM for 10 minutes at 2°C. The supernatant was subsequently removed and the RNA pellet was washed with 1.0 mL of 75% ethanol. After another centrifugation, the ethanol supernatant was removed and the pellet was dissolved in DEPC-treated water. Following a final incubation for 5 minutes at 55°C, the amount of RNA was quantified with a spectrophotometer. Use of the Affymetrix HT-HGU133A chips allowed all RNAs to be profiled within a single experiment, thereby minimizing batch effects and other technical concerns that might adversely affect data quality.

Tiling array CGH data processing The tiling array CGH was downloaded from the GEO database (GSE6779). Log2-ratios were calculated as follows: first, Cy3 and Cy5 channel values were determined. In the event that Cy3 (signal-background) fell below the median Cy3 background, the Cy3 (signal-background) was given the median Cy3 background value. This was similarly done for the Cy5 channel. Log2-ratios were obtained by dividing the Cy3 (signal - background) by the Cy5 (signal - background) divided by the (median Cy3 background divided by the Cy5 background). Missing values were replaced with the mean of the closest two neighboring markers. Finally, the log2-ratios were median-column normalized. Samples were then smoothed using the GLAD segmentation algorithm with default parameters. Markers in regions of known copy number polymorphisms were removed (http://projects.tcag.ca/variation/, October 2006) as in the other data sets.

SNP array and CGH data set comparisons Q-value distributions determined by GISTIC were used as a basis for quantitative comparison by Pearson’s-correlation coefficient. Since the various data sets employed had different numbers of markers, a composite marker set spanning the genome was first created from the combined data sets. Next, each q-value distribution was mapped to its original genomic location. Regions of the q-value distribution that were not reflected in the original data set were then filled in with neighboring original values. For the data set

2 comparisons, this process was performed on both q-value distributions, thus creating two distributions of equal size mapped by genomic position (see Supplementary Table 2). In our intra-data set comparison, a random number generator was used to assign individual samples to one of two subgroups. This randomization was repeated 100 times, GISTIC analysis was performed on each subgroup and the median Pearson’s correlation coefficient was taken. The Pearson’s correlation test was chosen to focus on the genomic landscape rather than the absolute q-value differences, thereby allowing fairer comparisons between data sets with different sample sizes and markers numbers. Similar trends were observed when the –log(q-value) or G score were used as the metric, as well as a distribution of medians and means of all samples at a particular marker. Source code was written in MATLAB R2006b.

Inferring loss-of-heterozygosity LOH calls were inferred on 86 normal samples (the same normal samples chosen as most similar to compute copy number) using dChipSNP software. This data was then used to infer LOH regions for tumor samples with a 0.5 inferred LOH call rate and removing LOH regions consistent with more than 5% of reference samples. The resulting data matrix was extracted from dChip and loaded into MATLAB, where the GISTIC algorithm was performed on the inferred LOH matrix.

Western blotting and biochemical studies Cells were lysed in TNN buffer containing a mini complete protease inhibitor tablet (Roche), 1 mM NaF, and 1 mM NaVO3. Lysate concentration was determined via Bradford assay; was then prepared in SDS gel loading buffer and boiled for 8 minutes at 95°C. Lysates of 60 µg of protein were resolved by denaturing gel electrophoresis using the BioRad Minigel system with 10% Tris-HCl gels (BioRad, Catalog Number: 161-1158). Protein was transferred to nitrocellulose membranes (80V for 80 minutes at 4°C) and probed using primary antibodies against p-Erk (anti-phospho- p44/42), total ERK (anti-p44/42), p-MEK (anti-phospho-MEK1/2 (Ser217/221)), total MEK (anti-MEK1/2) (Cell Signaling Technology), Cyclin D1 (sc-20044, Santa Cruz Biotechnology), and alpha-tubulin (anti-alpha-tubulin, Cell Signaling Technology). All primary antibodies were used at a 1:1000 dilution. The secondary antibody was anti- rabbit IgG, HRP-linked (Cell Signaling Technology) at a 1:1000 dilution. Biochemical MEK inhibition by CI-1040 was examined by treating sub-confluent melanoma cells with varying dilutions of CI-1040 for 24 hours. After harvest and protein preparation, Western blotting was performed with the p-ERK and total ERK antibodies described above.

3 Supplementary Figure Legends

Figure S1: GISTIC comparative plots of 250K SNP vs 50K SNP and tiling array data Top: statistically significant amplifications (A) (orange, red) and deletions (B) (light blue, dark blue) by GISTIC analysis, comparing the 50K XbaI (n=31) and 250K StyI (n=70) arrays, respectively. Bottom: amplifications (C) and deletions (D) for the Jonsson et al. tiling array CGH cell line data set (n=45) were compared to the 250K StyI array data set. Left axes show genomic position; top axes show FDR q-values and bottom axes show corresponding G-scores. FDR threshold = 0.25.

Figure S2: Genomic distinctions and similarities between melanoma clinical subtypes Significant amplifications (red) and deletions (blue) by GISTIC are shown for the Curtin et al. melanoma CGH data set(2), including 30 chronic sun induced samples (A), 40 non-chronic sun induced samples (B), 36 acral melanoma samples (C), and 20 mucosal melanoma samples (D). Left axes show genomic position; top axes show FDR q-values and bottom axes show corresponding G-scores. FDR threshold = 0.25. Known oncogene or tumor suppressor are labeled to right of the corresponding peak.

Figure S3: Chromosomal copy number changes at the BRAF and PTEN loci Heatmap views of the PTEN and BRAF locus with samples (columns) sorted by level of deletion (blue) or amplification (red) are shown for the Jonsson 33K tiling array CGH data set (n=45) ((A) and (D), respectively) and the 2K Curtin et al. BAC array CGH data set (n=70) ((B) and (E), respectively). The BRAF locus in the 250K and 50K data sets is also shown (C) (n=101). Genes at the respective loci are indicated (Y axis).

Figure S4: Phospho-ERK inhibition by the MEK inhibitor CI-1040 Immunoblots show p-ERK, total ERK, and cyclin D levels in selected melanoma short term cultures following incubation with increasing CI-1040 concentrations. The MALME- 3M melanoma cell line (MALME) is included as a control. Mutation status of BRAF and NRAS are denoted, along with the CI-1040 GI50.

Figure S5: Chromosomal correlates of MAP kinase dependency A whole genome copy number heatmap is shown (red=copy gains; blue=copy losses) for selected short term cultures sorted by CI-1040 GI50 values. BRAF and NRAS mutation status is indicated.

Figure S6: Matched expression and copy number for 10 outliers Gene expression values are plotted for CUL2 (A) and KLF6 (B), sorted by chromosomal hemizygous deletion or retention. These genes were identified by whole genome SAM analysis (Fig. 3D of main text).

4 Supplementary Table 1: BRAF, NRAS, and PTEN mutation status

HTA Gene Sample SNP array Sample Name Expression BRAF NRAS PTEN* Other Type platform data 501MEL Cell line 50K XbaI yes WT G12D no data A375 Cell line 50K XbaI yes V600E WT no data Del exon 2; HS944 Cell line 250K StyI no no data Q61K** frameshift/premature stop** Short term KO13 50K XbaI no no data no data no data culture Short term KO16 50K XbaI no V600E WT WT culture Short term KO18 50K XbaI no WT WT WT culture Short term KO25 50K XbaI no no data no data no data culture Short term KO27 50K XbaI no V600E WT no data culture Short term PIK3CA KO28 50K XbaI no V600E WT WT culture E542K Short term KO29AX 50K XbaI no V600E WT WT culture Short term KO33b 50K XbaI no no data no data no data culture Short term KOO6 50K XbaI no WT WT no data culture Short term KOO8 50K XbaI no V600E WT WT culture LOX_IMVi Cell line 50K XbaI no V600E WT no data Short term M000216 250K StyI yes V600E WT WT culture Short term M000301 250K StyI yes WT Q61K WT culture Short term M000907 - yes WT G13R WT culture Short term M000921 250K StyI yes V600E WT T319fs*1 culture Short term M010124 250K StyI yes WT Q61K WT culture Short term M010308 250K StyI yes WT Q61K WT culture Short term M010322 - yes V600E WT WT culture Short term M010403 250K StyI yes no data no data WT culture Short term M010606 250K StyI yes WT WT WT culture Short term M010718 - yes WT Q61K WT culture Short term M010817 250K StyI yes WT Q61R P213L culture Short term M010920 250K StyI yes WT WT WT culture Short term M13 50K XbaI no WT Q61K WT culture M14 Cell line 50K XbaI no V600E no data no data Short term M16 50K XbaI no WT WT WT culture Short term M19 50K XbaI no no data no data no data culture Short term M23 50K XbaI no WT Q61K no data culture Short term M25 50K XbaI no V600E no data no data culture Short term M26 50K XbaI no WT WT WT culture

5 Short term M34 50K XbaI no V600E WT K164fs*16 culture Short term M970109 - yes V600E WT WT culture Short term M970131 - yes WT WT WT culture Short term M970805 - yes WT WT WT culture Short term M980409 - yes WT Q61K WT culture Short term M980513 - yes no data no data WT culture Short term M980928 250K StyI yes WT WT WT culture Short term M981201 - yes WT Q61K WT culture Short term M990114 250K StyI no WT WT WT culture Short term M990514 - yes no data no data WT culture Short term M990719 250K StyI yes WT G13R no data culture Short term M990802 250K StyI yes V600E WT WT culture Short term M991121 - yes WT WT WT culture MALME3M Cell line 50K XbaI yes V600E WT no data Breast MCF7 cancer cell - yes no data no data no data line MDAMB435 Cell line 50K XbaI no V600E no data no data Short term MEL35 50K XbaI no WT Q61R no data culture Short term MEL9 50K XbaI no no data no data no data culture Normal MelanocyteW2 - yes no data no data no data melanocyte Normal MelanocyteW3 - yes no data no data no data melanocyte Normal MelanocyteW4 - yes no data no data no data melanocyte Normal MelanocyteW6 - yes no data no data no data melanocyte Normal MelanocyteW7 - yes no data no data no data melanocyte MeWo Cell line 250K StyI no WT* no data no data Short term ML310 - yes no data no data no data culture SK-MEL-119 Cell line 250K StyI no no data Q61R** WT* SK-MEL-19 Cell line 250K StyI no V600E** no data no data SK-MEL-2 Cell line 50K XbaI yes WT Q61R no data SKMEL28 Cell line 50K XbaI yes V600E WT T167A* CDK4 R24C SK-MEL-30 Cell line 250K StyI no no data Q61K** WT* SKMEL5 Cell line 50K XbaI no V600E WT no data SK-MEL-63 Cell line 250K StyI no no data Q61K** WT* UACC257 Cell line 50K XbaI no V600E WT no data UACC62 Cell line 50K XbaI no V600E WT no data Short term WM1026 250K StyI yes WT WT WT culture Short term WM1325 250K StyI yes no data no data no data culture Short term WM1346 250K StyI yes no data no data no data culture Short term WM1385 250K StyI yes no data no data no data culture

6 Short term WM1419 250K StyI yes no data no data no data culture Short term WM1433 250K StyI yes no data no data no data culture Short term WM1480 250K StyI yes no data no data no data culture Short term WM1575 - yes no data no data no data culture Short term WM1716 250K StyI yes V600E WT no data culture Short term WM1719 - yes no data no data no data culture Short term WM1720 250K StyI yes V600E WT WT culture Short term WM1745 250K StyI yes V600E WT no data culture Short term WM1852 250K StyI yes V600E WT WT culture Short term WM1862 250K StyI yes V600E WT no data culture Short term WM1930 250K StyI yes V600E WT WT culture Short term WM1931 250K StyI no V600E WT WT culture Short term WM1942 - yes V600E WT WT culture Short term WM1960 250K StyI yes WT Q61K WT culture Short term WM1963 - yes WT WT WT culture Short term WM1968 250K StyI yes no data no data WT culture Short term WM1976 250K StyI yes V600E WT WT culture Short term WM2029 250K StyI yes WT WT WT culture Short term WM262 250K StyI yes no data no data no data culture Short term WM3061 250K StyI yes WT G13R WT culture Short term WM3066 - yes no data no data no data culture Short term WM3077 250K StyI yes no data no data no data culture Short term WM3130 250K StyI yes K601E WT WT culture Short term WM3163 250K StyI yes V600E WT WT culture Short term WM3215 250K StyI yes V600E WT T319fs*1 culture Short term WM3218 250K StyI yes V600E WT WT culture Short term WM3228 250K StyI yes V600E WT WT culture Short term WM3243NCI-A5 250K StyI yes V600E WT WT culture Short term WM3244NCI-B - yes no data no data WT culture Short term WM3246 250K StyI yes WT WT K80_R84del culture Short term WT (homozygous WM3259 250K StyI yes V600E WT culture deletion) Short term WM3282 250K StyI yes V600K WT WT culture Short term WM3297 250K StyI yes V600E WT WT culture Short term WM3311 250K StyI yes WT WT WT culture Short term WM3314 - yes no data no data no data culture

7 Short term WM3381A 250K StyI yes V600E WT no data culture Short term WM3450 250K StyI no WT Q61K WT culture Short term WM3451 - yes no data no data WT culture Short term WM3456 250K StyI yes WT Q61K WT culture Short term splice+2 insG, WM3457 250K StyI yes V600E WT culture following aa267 Short term WM3482 250K StyI yes V600E WT WT culture Short term WM3506 250K StyI yes WT Q61R WT culture Short term WM3526 250K StyI yes V600E WT WT culture Short term WM3619 250K StyI yes WT Q61R WT culture Short term WM3623 250K StyI yes WT Q61K WT culture Short term WM3627 250K StyI yes V600E WT WT culture Short term WM3629 250K StyI yes D594G G12D WT culture Short term WM3670 250K StyI yes G469E G12D WT culture Short term FGFR1 WM3682 - yes WT WT WT culture S125L Short term FGFR1 WM3702 250K StyI yes WT WT no data culture S125L Short term WM3727 250K StyI yes V600E WT WT culture Short term WM451Lu 250K StyI yes V600E WT WT culture Short term WM46 - yes no data no data no data culture Short term WM806 250K StyI no no data no data no data culture Short term WM853-2 250K StyI yes V600E WT WT culture Short term WM858 250K StyI yes no data no data no data culture Short term WM984 250K StyI yes V600E WT WT culture WW94 Cell line 50K XbaI yes no data no data no data *PTEN mutation nomenclature follows COSMIC with the exception of sample WM3457 **previously reported (18)

8 Supplementary Table 2: Genomic diversity of melanoma short term cultures

Data set 1 Data set 2 Pearson Correlation Coefficient cultured (a) cultured (b) 0.85 noncsd + csd cultured 0.71 noncsd cultured 0.66 mucosal cultured 0.55 csd cultured 0.52 acral cultured 0.48 lung noncsd + csd 0.42 lung cultured 0.35 normal cultured 0.016 normal lung 0.02 normal noncsd + csd 0.01 normal jonsson 0.01 jonsson cultured 0.80 jonsson noncsd + csd 0.74 jonsson acral 0.49 jonsson mucosal 0.56 jonsson csd 0.47 jonsson noncsd 0.69 cultured short term cultures and cell lines (n=89, 250K StyI and 50K Xba) (a) and (b) were randomly selected mutually exclusive subsets from 89 (see Supplementary Methods) csd primary chronic sun-induced damage melanoma (n=30, 2K BAC array CGH, Curtin et al.) noncsd primary non chronic sun-induced damage melanomas (n=40, 2K BAC array CGH, Curtin et al.) csd + noncsd cutaneous primary melanomas (csd and noncsd, n=70, 2K BAC array CGH, Curtin et al.) acral primary acral melanoma (n=36, 2K BAC array CGH, Curtin et al.) mucosal primary mucosal melanoma (n=20, 2K BAC array CGH, Curtin et al.) lung primary lung adenocarcinomas (n=462, 250K Sty I, Weir et al., submitted) normal normal tissue (n=156, 250K StyI) jonsson melanoma cell lines (n=45, 30K tiling array CGH, Jonsson et al.)

9 References

1. Hupe P, Stransky N, Thiery JP, Radvanyi F, Barillot E. Analysis of array CGH data: from signal ratio to gain and loss of DNA regions. Bioinformatics 2004;20:3413-3422. 2. Curtin JA, Fridlyand J, Kageshita T, et al. Distinct sets of genetic alterations in melanoma. N Engl J Med 2005;353:2135-2147.

10 Supplementary Figure 1 - Lin et al.

FDR q-values FDR q-values 1.0e-9 1.4e-15 A 1.0e-09 1.4e-15 7.6e-25 B 3.0e-2 1.0e-3 7.6e-25 1.5e-39 8.0e-63 3.1e-2 1.0e-3 4.9e-6 1 0.25 4.9e-6 0.25

1

1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 9

Chromosomes 10 10 11 11 12 12 13 13 14 14 15 15 16 16 17 17 18 19 18 20 19 20 21 22 21 22 0.10 0.12 0.16 0.19 0.24 0.32 0.34 0.14 0.18 0.22 0.28 0.40 0.53 0.66 0.92 1.10

G score G score

FDR q-values FDR q-values 1.0e-9 1.0e-9 C 1.4e-15 D 1.4e-15 1.0e-3 7.6e-25 1.0e-3 7.6e-25 1.5e-39 8.0e-63 1 0.25 0.031 4.9e-6 1 0.25 0.031 4.9e-6

1 1 2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 9 10 10 11 11 12 12 13 13 14 14 15 15 16 16 17 17 18 18 19 19 21 20 20 22 21 22 0.08 0.10 0.12 0.16 0.19 0.24 0.32 0.34 0.14 0.18 0.22 0.28 0.40 0.53 0.66 0.92 1.10

G score G score Supplementary Figure 2 - Lin et al.

A Chronic sun induced damage B Non chronic sun-induced damage

Amplifications Deletions Amplifications Deletions FDR q-value FDR q-value FDR q-value FDR q-value

1.0 0.0068 0.0010 6.9e-3 1.0e-3 6.9e-3 0.55 0.25 0.10 0.03 9.7 1 0.55 0.25 0.10 0.031 9.7 4.9 1 0.55 0.25 0.10 0.031 1.0e-3 9.7 1 0.55 0.25 0.10 0.031 6.9e-3 1.0e-3 9.7e-5 4.9e-6 1.1e-7

e e e e -5 -5 -6 -5 e -9

1 1 1 1 2 2 2 2 3 3 3 3 4 4 4 4

5 5 5 5 6 6 6 6 7 7 7 7

8 8 8 MYC 8 CDKN2a CDKN2a omosome omosome 9 9 9 9 r r 10 10 10 10 ch ch

11 CCND1 11 11 CCND1 11 12 12 12 12 13 13 ERCC5? 13 13 ERCC5? 14 14 14 14 15 15 15 15 16 16 16 16 17 17 17 17 18 18 18 18 19 19 19 19 20 20 20 20 21 21 21 21 22 0.029 0.055 0.070 22 0.0519 0.0720 0.0845 0.097 22 0.045 0.059 0.074 22 0.042 0.050 0.061 0.071 0 0.085 0.118 0.150 0.173 0.199 0.034 0.114 0.133 0.146 0.188 0.031 0.099 0.123 0.165 0.083 .094 0.108 0.121 0.144 0.165 0.199

G-score G-score G-score G-score

C Acral D Mucosal Amplifications Deletions Amplifications Deletions FDR q-value FDR q-value FDR q-value FDR q-value 1.0 1.0 1.4 7.6 1.4 7.6 1 0.25 0.031 1.0e-3 4.9 1 0.25 0.031 1.0e-3 4.9 0.55 0.25 0.10 0.031 6.9e-3 1.0e-3 9.7 0.55 0.25 0.10 0.031 6.9e-3 1.0e-3 9.7

e e e e e e e e -6 e -15 -25 -6 e -15 -25 -5 -5 -9 -9

1 1 1 1

2 2 2 2 3 3 3 3

4 4 4 4

5 5 5 5 6 6 6 6 7 7 7 7

8 MYC 8 8 8 MYC CDKN2a 9 9 CDKN2a 9 9 omosome omosome r r 10 10 10 10 ch ch

11 CCND1 11 11 11 12 CDK4 12 12 CDK4 12 MDM2 MDM2 13 13 ERCC5? 13 13 14 14 14 14 15 15 15 15 16 16 16 16 17 17 17 17 18 18 18 18 19 19 19 19 20 20 20 20 21 22 21 22 21 22 21 22 0.068 0.133 0.187 0.274 0.388 0.536 0.646 0.054 0.080 0.109 0.143 0.177 0.231 0.274 0.104 0.137 0.170 0.203 0.248 0.320 0.386 0.489 0.090 .118 0.149 0.183 .227 0.268 0.304 0.329 0 0 G-score G-score G-score G-score Supplementary Figure 3 - Lin et al.

A B q21. SFTPD ZWINT ANXA11 TSPAN14 IPMK PHYHIPL ANK3 q21.2TMEM26 ZNF365 C10orf74

q21.3 NRG3 CTNNA3 DNAJC12 q23. CCAR1 COL13A1 q22.SGPL1 C10orf54 GHITM C10orf42 FLJ32658 KIAA1128 q22.2MYST4 C10orf11 KCNMA1

q22.3RAI17 GRID1 SFTPD q23.2 KIAA0261 NRG3 q23. BMPR1A GHITM GLUD1 FAM22D q23.2GRID1 PAPSS2 MMRN2

PTEN q23.3ANKRD22 C10orf59 PTEN MPHOSP LIPF HTR7 AMSH-LP q23.3TNKS2 q23.3CH25H SEC15L1 q23.3 SLC16A12 TMEM20 MPHOSP CYP2C9 C10orf13 q24. MLR2 C10orf13 q24.2HPSE2 HTR7 ENTPD7 q24.3PAX2 PCGF5 q24.3FGF8 PPP1R3C ARL3 q23.3 q24.3 C10orf13 SH3MD1 CPEB3 SORCS3 IDE q25.SORCS1 SEC15L1

FER1L3 XPNPEP1 PDE6C PDCD4 q23.3 q25.2 TMEM20 GPAM PLCE1 TCF7L2 ADRB1 HELLS q25.3TRUB1 CYP2C19 CYP2C8 GFRA1 PDLIM1 VAX1 q26.CASC2 ALDH18A NANOS1 C10orf13 q26.PPAPDC1 q24 DNTT FGFR2

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8

C D PLXNA4B CHCHD3

SEC8L1

TBXAS1 FLJ32786 AKR1B1 CALD1 PARP12 FLJ11000 q33 CNOT4 SLC13A4 MTPN

CHRM2 PTN SLC37A3 DGKI

RAB19B CREB3L2

MKRN1 TRIM24 ATP6V0A MGC1428 DENND2A HSPC268 HIPK2 TBXAS1 ADCK2 SLC37A3 ADCK2 BRAF q34 BRAF LOC4014 BRAF FLJ40852 q34 TAS2R38 TRY1

MRPS33 PRSS1 OR9A2 GSTK1 LOC4412 OR2F2 FLJ43692

TPK1

LOC4014 q35

CNTNAP2 MULK

FLJ40852 C7orf33 TAS2R3 EZH2 DKFZp76 LOC1362 KIAA1285

OR9A4

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8

E q31.2TES MET WNT2 CTTNBP2

q31.3

KCND2 FLJ21986 PTPRZ1 q31.3CADPS2 SLC13A1 ASB15

GPR37 q31.3

GRM8 LOC1688 SND1 q32.RBM28 FLNC KIAA0828 q32.2UBE2H CPA1

MKLN1 q32.3 PLXNA4B

SEC8L1 FLJ32786 BPGM FLJ11000 q33 SLC13A4

CHRM2 DGKI TRIM24 MGC1428 HIPK2 PARP12 DENND2A q34 LOC4014 BRAF OR9A4 PRSS1 TAS2R40 KIAA0738 OR2A1

q35

CNTNAP2

C7orf33 DKFZp76 FLJ12700

q36.MGC3036

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 Supplementary Figure 4 - Lin et al.

ERK V600E GI50= 0.06 uM

Cyclin D1

CI-1040 (uM) 0 0.016 0.08 0.4 2 10 CI-1040 (uM) 0 0.016 0.08 0.4 2 10

pERK pERK MALME WM853-2 ERK V600E, WT ERK V600E, WT BRAF V600E GI50= 0.06 uM GI50= 0.67 uM Gi50: 60 nM Cyclin D1 Cyclin D1

CI-1040 (uM) 0 0.016 0.08 0.4 2 10 CI-1040 (uM) 0 0.016 0.08 0.4 2 10 pERK pERK WM3629 WM3670 WT, G12D ERK G469E, G12D ERK GI50= 1.63 uM GI50= 0.36 uM BRAF G469E, CNRycASlin G12 D1D Gi50: 360 nM Cyclin D1

CI-1040 (uM) 0 0.016 0.08 0.4 2 10 CI-1040 (uM) 0 0.016 0.08 0.4 2 10

pERK pERK WM1960 WT, Q61K WM3130 ERK ERK GI50= 0.065 uM K601E, WT BRAF K601E GI50= 1.73 uM CGi50ycl:in 1 7D130 nM Cyclin D1 Supplementary Figure 5 - Lin et al.

CI-1040 0.06 0.06 0.36 GI50 0.023 0.023 0.035 0.037 0.045 0.046 0.048 0.065 0.106 0.124 0.151 0.215 0.316 0.339 0.447 0.485 0.673 1.632 1.732 3.461 PTEN HD WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT MUT MUT no data NRAS WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT WT Q61K Q61K Q61K Q61R G12D G12D G13R BRAF WT WT WT WT WT V600E V600E V600E V600E V600K V600E V600E V600E V600E V600E V600E V600E V600E V600E V600E K601E G469E D594G WM3457 WM3623 WM3482 WM3526 WM3228 WM3282 WM3506 WM1976 WM3627 WM1960 WM3163 WM3218 WM1862 WM3215 WM3259 WM3670 WM3456 WM1930 WM3629 WM3130 WM3061 WM853-2 WM451Lu 1 2 3 4 5 6 7 8 9 10 11 Chromosomes 12 13 14 15 16 17 18 19 21 20 22

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 Supplementary Figure 6 - Lin et al.

A B 9.5 11

9 10 9 8.5 8 8 7 (gene expression)

7.5 (gene expression) 2

2 6 log 7 log CUL2 5 KLF6 loss retention loss retention