Supplementary Materials for Contrasted sex chromosome evolution in primates with and without sexual dimorphism Rylan Shearn, Emilie Lecompte, Corinne Régis, Sylvain Mousset, Simon Penel, Guillaume Douay, Brigitte Crouau-Roy, Gabriel A.B. Marais Correspondence to: [email protected] This PDF file includes: Supplementary Text S1 to S2 Figs. S1 to S2 Tables S1 1 Supplementary Text Text S1: Regions of the strepsirrhine X chromosomes with unusual male:female coverage ratio In Fig. 1, both lemur X chromosomes exhibit regions with male:female coverage ratio close to 1 (shown in grey) in their X-specific parts, where a ratio of 0.5 is expected. The gray mouse lemur has five such regions, the northern greater galago three. The dot plots of the lemur and the human X chromosomes (see Fig. 1 and S1) clearly show that little or no homologous genes are found in those regions, which suggest that they may be homologous to other human chromosomes. This would be consistent with the male:female coverage ratio of 1, typical of autosmal regions, that we found for these regions. To explore this possibility, we extracted the sequences of those regions and performed a tblastn against all the human proteins (human genome version GRCh38). In case of isoforms, the longest protein was kept so that a human gene was present only once. We then filtered the tblastn results by keeping only hits with >80% similarity (based on average nucleotide divergence between lemurs and humans) and e-value < 10-9. From those, we kept human proteins covered by hits to >80% using SiLix (Miele, Penel, & Duret, 2011). Only proteins matching to no more than one region were kept. The results of the tblastn are shown in the table below. Microcebus murinus X chromosome regions* Otolemur garnetti X chromosome regions* Human 30.3- 41.6- 46.8-48 61.5- 92.7- 49.5- 80-84.5 116- chromosomes 33.2 44.1 63.7 93.7 68.5 133 Chrom. 1 4 54 13 2 1 Chrom. 2 2 4 1 Chrom. 3 1 Chrom. 4 2 Chrom. 5 2 2 1 Chrom. 6 10 Chrom. 7 2 Chrom. 8 1 1 4 1 1 1 Chrom. 9 1 Chrom. 10 Chrom. 11 Chrom. 12 8 1 2 3 15 Chrom. 13 1 1 44 Chrom. 14 4 1 1 Chrom. 15 2 2 Chrom. 16 1 1 1 Chrom. 17 1 3 1 Chrom. 18 2 Chrom. 19 1 3 Chrom. 20 119 Chrom. 21 Chrom. 22 1 1 Chrom. X 2 1 2 3 *coordinates in Mb Human chromosome with the largest number of homologs is shown in bold 2 For all regions except one, most homologs that we identified are from the human autosomes, which confirms our hypothesis. These homologs are mainly from one source: chromosomes 1, 8 and 12 for regions 46.8-48, 61.5-63.7, 92.7-93.7 and 41.6-44.1 of the gray mouse lemur X chromosome, and chromosomes 12, 13 and 20 for regions 80-84.5, 116-133 and 49.5-68.5 of the northern greater galago X chromosome. These results can be interpreted two ways. A possibility is that the assemblies of those lemur X chromosomes wrongly include autosomal scaffolds. Another possibility is that during the evolution of strepirrhines, some autosomal fragments have been translocated to the PAR, and the assembly failed to order these fragments correctly. Our approach cannot tell apart these possibilities but in all cases, our results suggest that these regions are probably assembly errors. Changing tblastn outputs filtering did not change qualitatively the results. With lower %identity thresholds, we detected autosomal homologs for region 30.3-33.2 (for example, with %identity > 65, we found 2 proteins from chrom. 1, 1 from chrom. 2 and 1 from chrom 19). References Miele, V., Penel, S., & Duret, L. (2011). Ultra-fast sequence clustering from similarity networks with SiLiX. BMC Bioinformatics, 12, 116. doi: 10.1186/1471-2105-12-116 3 Text S2: Statistical tests on strata formation Assuming a constant rate l for the formation of new evolutionary strata, the number S of new strata formed during a time interval Dt is Poisson-distributed with parameter lDt k lDt (lDt) e− P(S = k)= . (1) k! We partition the time interval Dt into two parts Dt = Dt1 +Dt2. During the time interval Dti we observe the formation of Si new strata. We want to contrast the following two hypotheses: • H0: Strata accumulated at a common rate l0. • H1: Strata accumulated at different rates li during the time intervals Dti. Likelihood ratio test Differentiating equation (1) with respect to l enables to find the maximum likelihood estimator for l: ˆ S ˆ S1 + S2 ˆ Si l = l0 = andli = . (2) Dt ) Dt1 + Dt2 Dti A likelihood-ratio test can be used to compare the likelihoods L0 and L1 of the models underlying the null and alternative hypotheses H0 and H1. ˆ L0 = P S = k1 + k2 l = l0,Dt = Dt1 + Dt2 , | ⇣ ˆ ⌘ ˆ L1 = P S1 = k1 l1 = l1,Dt = Dt1 P S2 = k2 l2 = l2,Dt = Dt2 . | ⇥ | ⇣ ⌘ ⇣ ⌘ L0 2 Under the null hypothesis H0, the likelihood-ratio statistics X = 2log is approximately c -distributed − L1 with 1 degree of freedom. Exact binomial test An alternative method consists in using the number S1 of strata formed in the time interval Dt1 as the test statistics and compute the the conditional probability to observe a larger value given the total number of strata formed in the time interval Dt1 + Dt2: k +k k +k j k1+k2 j ( t + t ) (k + k )! 1 2 l 1 2 Dt Dt − e l D 1 D 2 (S k k + k , t , t )= 1 2 ⇥ 1 ⇥ 2 ⇥ − ⇥ P 1 > 1 1 2 D 1 D 2 k +k l (Dt +Dt ) Â | (l (Dt1 + Dt2)) 1 2 e 1 2 j!(k1 + k2 j)! ⇥ − ⇥ j=k1 − k +k j k1+k2 j 1 2 k + k Dt Dt − = Â 1 2 1 2 , j Dt1 + Dt2 Dt1 + Dt2 j=k1 ✓ ◆✓ ◆ ✓ ◆ (3) where we recognize the binomial distribution. Note that this is now independent of the rate l of the Poisson process. Basically, applying this test is conceptually equivalent to tossing an unbalanced coin k +k times with a probability p = Dt1 to get a head and computing the probability to obtain at least 1 2 Dt1+Dt2 k1 times a head. Evolutionary times The phylogenetic relationships and mean divergence times for the included primate species were recovered from the previously published primate phylogeny and divergence dates (Pozzi et al., 2014, supplementary table 3). Detailed phylogenetic relationships among strepsirrhine lineages (Horvath et al., 2008) were used to infer phylogenetic reletionships in the cases when species in our analysis were not 1 4 included in this reference study. The values are reported in the following table. Node labela Mean agea (My) Connected nodes or species 4 7.65 Homo sapiens Pan troglodytes 5 10.63 4 Gorilla gorilla 7 17.29 5 Pongo pygmaeus 38 32.12 7 Macaca mulatta 44b 46.72 38 Callithrix jacchus 61 74.11 44 60 50c 24.24 Eulemur rufus Hapalemur simus 54d 43.46 50 Microcebus murinus 55 59.55 54 Daubentonia madagascarensis 56 17.28 Galago senegalensis Otolemur garnetti 58 36.35 56 Nycticebus cougang 60 66.33 55 58 aFor sake of clarity we report the node labels and ages from our reference primate phylogeny (Pozzi et al., 2014). bCallithrix jacchus is not included in the reference phyolgeny but the divergence time between platyrrhini and catarrhini can be used. cNeither Eulemur rufus nor Hapalemur simus are included in the reference phylogeny, but the divergence between Eulemur macaco and Lemur catta can be used (node 18 in Horvath et al., 2008). dMicrocebus murinus is not included in the reference phylogeny, but the divergence with Lepilemur sp. can be used (node 7 in Horvath et al., 2008). Results The catarrhini and platyrrhini lineages evolved for Dt1 = 188.52 My during which S1 = 3 new strata were formed. The strepsirrhini lineages evolved for Dt2 = 321.32 My and no new strata was formed (S2 = 0). A likelihood-ratio test comparing these rates would be significant (X = 5.95, p = 0.015) however so few observations certainly violate the conditions for the convergence of the null distribution to the c2 distribution. The result of the exact binomial test approaches statistical significance (one-tailed test, p = 0.0506), suggesting new strata could form at a higher rate in the catarrhini and platyrrhini than in the strepsirrhini lineages. One has to keep in mind however that we used mean divergence times. Although 95% confidence intervals are available for the divergence times, we are not able to provide confidence intervals for the durations Dt1 and Dt2. Moreover, it would make sense to scale durations based on the generation times. Our approach assumes equal average generation time but shorter generations in the strepsirrhini lineages would lead to a more significant difference. References Horvath, J. E., D. W. Weisrock, S. L. Embry, I. Fiorentino, J. P. Balhoff, P. Kappeler, G. A. Wray, H. F. Willard, and A. D. Yoder. 2008, Development and application of a phylogenomic toolkit: resolving the evolutionary history of madagascar’s lemurs. Genome research 18:489–499. Pozzi, L., J. A. Hodgson, A. S. Burrell, K. N. Sterner, R. L. Raaum, and T. R. Disotell. 2014, Pri- mate phylogenetic relationships and divergence dates inferred from complete mitochondrial genomes. Molecular phylogenetics and evolution 75:165–183. 5 2 GG GGGGG PAR2 GGGG GGG GGG GGG GGG 150 G G G G GGG GG GG GGG GGG GGG G G GGG GGG GG GGG GGG G G G GGG GGG GGG GGGG GGG G G GGG G GGG G S1 GGG GGG GG GG G G GGG GGGGG GGGGGG G GGGGG 100 GGG GGG Region G misA G G mPAR G GG GG G PAR1 G G GG GGGG PAR2 GG GG GGG humanX G S1 G G GGGG GG GGG G G GG GG S23 S4 S5 G GGG GGG GG GGG 50 GG GGG GGG GGG GG GG G GGG GG S2 G S3 G G GGG G G G G G G G G G G GGG G G GGG GG GGG G GG GG GG GG GGG GG GG GG GG GG GGG GG G S4 GG S5 GG GGG G PAR1 GG 0 G 2.0 1.5 Region 1.0 misA mPAR 0.5 Otolemur (M/F) Otolemur 0.0 0 50 100 150 Otolemur Chr X (Mbp) Fig.