Evolutionary forces at work in the human genome

Conservation of genomic sequences through evolutionary history provides a powerful tool to guide biological investigations towards functionally important regions. In this work I have investigated two sets of such conserved sequences. The first set are the Ultraconserved Elements (UCEs). These are stretches of at least 200 basepairs (bp) of DNA in the human genome that match identically with corresponding regions in the mouse and rat genomes. Most UCEs are noncoding and have been evolutionarily conserved since mammal and bird ancestors diverged over 300 million years ago. The second set are the Human Accelerated Regions (HARs). These are elements ranging in size from 100--400bp that were generally conserved throughout vertebrate evolution, but that show an unexpected number of human-specific changes. For both analyses I used human polymorphism data. For the UCEs, resequencing of genomic DNA from 72 diploid human samples was performed by the sequencing center at Washington University using conventional capillary electrophoresis (CE) methods. I aggregated the small number of polymorphic sites found in more than 300 UCEs and analyzed the resulting derived allele frequency (DAF) spectrum to show that the UCEs as a whole are under negative (purifying) selection that is much stronger than that in protein coding genes. In contrast, my analysis of the top 49 HARs used the latest generation of high throughput sequencing technology. I implemented a technique to enrich the genomic DNA from 11 human samples (i.e. 22 chromosomal samples per locus) for 40kb neighborhoods of these 49 HARs. After obtaining short read sequences for this enriched DNA from the sequencing center at UC Santa Cruz, I mapped these reads to the reference human genome, used the mapped sequences to call polymorphic genotypes, and constructed the DAF spectrum for each HAR. My analysis of these spectra showed only weak evidence for selective forces, which was not significant after correction for multiple hypothesis testing. However, I performed other analyses considering the pattern of nucleotide substitutions from the ancestral to the derived alleles. I found that in addition to an historical bias favoring the conversion of weak (A or T) alleles into strong (G or C) alleles there is also strong evidence in the DAF spectra of many of these 40kb regions for ongoing weak-to-strong (W2S) fixation bias. Together, these results do not rule out functional roles for the observed changes in the HARs---indeed there is good evidence that the first two HARs are functional elements in humans---but they suggest that a fixation process (such as biased gene conversion) that is biased at the nucleotide level but is otherwise selectively neutral, could be an important evolutionary force at play in them, both historically and at present.