Positive selection on cis -regulatory sequences during human evolution R. Haygood, O. Fedrigo, B. Hanson, J. Pavisic, K. Yokoyama, and G. A. Wray Biology Department and Institute for Genome Sciences and Policy, Duke University
Genetic differences between humans and chimpanzees that arose through positive selection on the human lineage • Evolutionarily interesting and medically relevant • Many examples known in protein-coding sequences • Few examples known in cis -regulatory sequences Yet cis -regulatory sequences may be crucial.
A genome-scale scan of promoter regions • For faster evolution along the human lineage • In regions immediately upstream from transcript start sites (5’ flanking regions) • Than of nearby predominantly neutral intronic sequences (non-first introns minus splicing signals) • Using the publicly available human, chimpanzee, and macaque genome sequences
Maximum-likelihood fitting • Of a pair of substitution models, one null and the other alt • Involving ratio ζ of rates in upstream region to those of intronic sequences (analogous to d N /d S for coding sequences) • null constrains ζ ≤ 1 (negative or no selection), alt allows ζ > 1 (positive selection) on human lineage only • Likelihood-ratio test indicates whether alt fits data significantly better than null
Quality control • Stringent filtering to eliminate ambiguous or low-quality sequences and alignments • Replicated fitting of each model to each gene • Bootstrapping over intronic sequences Downstream regions where the next gene downstream is transcribed in the opposite direction yield a weaker signal.
Vital statistics • ~6300 genes can be analyzed using intronic A (~7800 using intronic B) • alt 1 vs. null 1 and alt 2 vs. null 2 yield 15 and 26 genes with q -values < 0.05 using intronic A (25 and 32 using intronic B) • Correlation between p -values for alt 1 vs. null 1 and those for alt 2 vs. null 2 is 0.86 using intronic A (0.87 using intronic B)
PANTHER biological process categories category # genes MW p 10% p mRNA transcription elongation 10 0.0014 0.013 Proteolysis 322 0.022 0.0027 Protein metabolism and modification 882 – 0.0041 Anion transport 32 0.025 0.012 Monosaccharide metabolism 23 0.042 – Protein complex assembly 23 0.050 – Macrophage-mediated immunity 47 – 0.041 Antioxidation and free radical removal 14 – 0.044
Immunity • Many known immunity-related genes score high, some on human lineage only (e.g., CD69, IL12A), others on chimp lineage too (e.g., HLA-DRB1, CD58) • High-scoring genes in 26S proteasome pathway (MW p = 0.0026, 10% p = 0.00050) may well be immunity-related (e.g., PSMC3, PSMC6) • Relevant to differences between humans and chimps in susceptibility to diseases such as malaria and AIDS?
Nutrition • Many known nutrition-related (i.e., digestion, metabolism, excretion) genes score high, mostly on human lineage (e.g., PPAP2C, NPHS1, KL) • Some high-scoring genes in Anion transport category (MW p = 0.025, 10% p = 0.012) are nutrition-related (e.g., SLC22A8, SLCO4C1, CLCNKB) • Reflects extensive evolution of human diet?
Cognition • Some known neural-related (i.e., neuronal function, axon guidance, brain development) genes score high (e.g., NPAS1, ISL2, IMPA2) • Other high-scoring genes in the Anion transport category are neural-related (e.g., GABRG1, GLRA1, GABRA4) • Mutations in several such genes are associated with cognitive deficits (e.g., GABRA4 with autism, IMPA2 with bipolar disorder)
In progress • Similar analyses of UTRs • “Bayes Empirical Bayes” analyses characterizing contributions of individual nucleotides to regional signals of positive selection • Reporter assays measuring impacts of sequence differences on gene expression • Sequencing of orangutan and other primates Each additional species will appreciably increase the power of comparative genomic analyses.
Thanks Advice Jim Kent Sergei Kosakovsky Pond Webb Miller Money Duke Institute for Genome Sciences and Policy National Science Foundation
Recommend
More recommend
