Did Dodging Disease Keep Pygmies Small? Patterns of Ancestry, Signatures of Natural Selection, and Genetic Association with Stature in Western African Pygmies PLoS GENETICS Joseph P. Jarvis, Laura B. Scheinfeldt., Sameer Soi., Charla Lambert., Larsson Omberg., Bart Ferwerda, Alain Froment, Jean-Marie Bodo, William Beggs, Gabriel Hoffman, Jason Mezey, Sarah A. Tishkoff* Pygmy: any group in whom adult males are under 1.52 meters live and hunt in tropical forests, and they die young (17~25y); infectious diseases WHY are they small/short? evolution Move nimbly through dense forest Fast heat-dissipating Trade-off between growth and reproduction before and after puberty etc. growth immunity converge Short stature is a byproduct!! Brief history of Pygmy ~60-70 kya African Pygmies Eastern Pygmies West African Bantu groups ~10-27 kya African mtDNA Western Pygmies ~4-5 kya Admixture Previous studies documented human growth hormone (HGH) and IGF-1 contribute to short stature of Pygmies. Materials: Western Pygmy (Baka, Bakola, and Bedzan): 67 (57 have phenotype data for height) Bantu-speaking (Tikar, Ngumba, and Lemande): 58 (39 have phenotype data for height) Platform: Illumina 1M SNP array Population Structure Proportion of Bantu ancestry Bakola: 27% Baka: 35% Bedzan: 49% Correlation between ancestry proportion and height Local Ancestry Inference Challenge to common methods: 1. The admixture is both ancient and on-going, thus a wide distribution of haplotype sizes including many that are relatively short is considered. 2. Only one of the two ancestral populations is still available and sampled (Bantu). Method: SupportMix SupportMix is a two-step process that classifies the ancestral origin of small regions of the genome using a support vector machine (SVM) trained on the ancestral population(s). This is followed by a classification of ancestry with the aid of a Markov model. The SVM learns the parameters of the Markov-process and the most probable ancestral state is solved by the forward-backward algorithm over the observed states. LAI results the average ancestry block sizes in Pygmies to be 1.7+/22.4 Mb for Pygmy ancestry, and 3.1+/24.6 Mb for Bantu ancestry A region with reduced levels of switching between ancestry blocks. Chr3: 46Mb-53Mb Scan for Selection: Fst/LSBL analysis; XP-EHH and his Analyses • Using samples estimated to have the lowest levels of admixture in the STRUCTURE analysis. Locus-Specific Branch Length (LSBL): diversity of each SNP Using three-way LSBL analysis helps to find the lineage specific differentiated SNPs • Adding a new population: Hapmap Maasai • Identifying SNPs that are differentiated specifically on the Pygmy lineage Lineage specific X-axis: x/(x+y+z) Y-axia: x high signal regions Fst > 0.33, (the top 0.1 percentile) 947 SNPs LSBL > 0.2215, (the top 0.1 percentile) 912 SNPs Most striking signal, Chr3: 45-60 Mb Several top signals from XP-EHH and iHS were also found in region Chr3: 45-60 Mb Visualization of high density blocks Pathway Enrichment Analysis of Scans for Positive Selection Method: PANTHER Region: all protein coding genes within 100kb up- and downstream of SNPs showing signatures of selection. Neuro-endocrine signaling and pathways potentially involved in reproduction and thyroid (甲状腺) function, including oxytocin (催产素) receptor mediated signaling and thyrotropin (甲状腺刺激激素)-releasing hormone receptor (TRHR) signaling. (genes near high Pygmy LSBL SNPs) • Oxytocin plays a role in lactation, parturition (分娩), and pair bonding. • TRHR signaling pathway plays a role in regulation of the hypothalamicpituitary-thyroid (下丘脑-脑垂体-甲状腺) axis and influences growth, thermo-regulation, reproduction, and immune response. Signatures of Selection and Association at GWAS Significant SNPs for Height in Non-Africans • 112 (of 211) SNPs significantly associated with height variation in recent GWAS in non-Africans are tested. • However, little evidence for selection at these loci were re-found. • Then, comparing the positional candidate genes reported in non-African GWAS to a list of all genes within 100 kb windows up- and downstream of signatures of selection in this study. • 69 overlapped, including DOCK3. • However, this overlap may be by chance • These results suggest that the genetic architecture of height in Pygmies differs from that observed in Europeans. Targeted Association Analysis for Height • First in Pygmy individuals only, 2288 SNPs. • Several suggestive associations with height, though not significant after multiple tests correction. • Next, combining Pygmy and Bantu dataset, for SNPs showing Pygmy-specific signatures of selection. • 128 SNPs showing significant associations after correction for multiple tests and for population structure. • Several genes found to be involved, say NDUFA4, DOCK3. • Multiple signals were again found in the region on Chr3: 45-60Mb. Pathway Enrichment Analysis in Extreme Genome-Wide Association Signals • Pathway enrichment for genes 100 kb up- and downstream of markers in the extreme tail of the empirical distribution of genome-wide p-values (the lowest 0.1%) • Top two pathways that are significantly enriched in the combined PygmyBantu association analysis are the protein kinase B signaling and the mitogen activated protein kinase/MAP kinase cascades of the Insulin/IGF pathway. Conclusions on the region of Chromosome 3: 45-60 Mb Three promising candidates: DOCK3, MAPKAP3, and CISH. CISH plays an import role in infectious disease susceptibility, and it also directly inhibits HGHR action and its expression is highly regulated by levels of HGH expression. Transgenic mice that overexpress CISH show reduced growth and overall small body size. Showing overlapping pattern of selection and association signals Signatures of selection in other population were also found in this region, which indicates its adaptation in human populations. Summary It is possible that the adaptive process that produced small body size in Pygmies may be the result of selection for traits other than stature, including early reproduction, metabolism, and immunity, and that the functional variants that are targets of selection may have pleiotropic effects. It is also possible that there have been multiple selective events in the history of the pygmy populations, at different time periods, that may have contributed to the co-adaptive evolution of loci that play a role in immunity, metabolism, and neuro-endocrine function.