FUNCTIONAL IMPACT OF POLYMORPHIC INVERSIONS IN THE HUMAN GENOME.

Abstract

An inversion is a balanced genomic rearrangement that alters the orientation of a specific genomic sequence. Despite not usually causing gain or loss of DNA, inversions can alter the original genetic background and produce mutational and positional effects on genes. In addition, inversions can alter recombination patterns both on the DNA sequences encompassed by them and in their vicinity. Therefore, inversions may associate with certain phenotypes or diseases, shape the evolutionary fate of the carriers by adaptive processes or even play a role in the origin of new species. For all that, increasing our knowledge of inversions constitutes a key issue for in-depth understanding of genome variation and its consequences. However, little is known about the prevalence and functional impact of inversions in the human genome, and in particular of their association with gene expression changes in humans. _x000D_ For the last 10 years, the advent of novel genomic technologies has enabled the study of SVs, including inversions, in a high-throughput fashion within and across species. State of the art genomic research provides the means and tools to explore the human genome in detail and expand our knowledge of inversions. Against this backdrop, the INVFEST project has devoted efforts to build the most accurate and exhaustive catalogue of human polymorphic inversions to date. For that, we have benchmarked GRIAL, currently the only paired-end mapping (PEM) based algorithm specifically designed to predict inversions, and demonstrated that performs with more accuracy and efficiency compared to other PEM based SV-detection methods, particularly in refining inversion BPs. In addition, we have curated GRIAL inversion predictions by using several pre and post PEM mapping filters coupled with manual inspection of complex cases and minimized the rate of false positive predictions._x000D_ Next, to gain further insight on the functional impact of polymorphic inversions in the human genome, we examined the overlap of 44 different inversions with genes. Overall, our results show that inversions tend to be located in intergenic regions but in 13.6% of the cases gene exons are affected. We performed two complementary approaches to identify inversion associations with gene expression. First, we performed a linear regression analysis in three different expression datasets derived from 527 lymphoblastoid cell lines (LCLs) of European, Asian and African HapMap individuals. Second, we interrogated blood and non-blood tissues by contrasting expression quantitative trait loci (eQTL) data with inversion tag-SNPs. We report 19 inversion rearrangements that modulate gene expression of 43 genes in several tissues. These results seem consistent as a subset (N = 11) of the associations found in LCLs has been identified by both approaches. Interestingly, we have identified an inversion that affects the expression of paralogous protein-coding genes (IFITM2/IFITM3) in lymphocyte-derived cells. We have also validated the methodology used for differential expression analysis by reproducing known effects of two well-studied inversions (17q21.31, 8p23.1) and identified novel associations of both inversion haplotypes with the expression of 6 genes. Moreover, we observe that at least 2 of these genes are associated to 17q21.31 structural rearrangements. Finally, we have looked for possible associations of inversions with disease and found one inversion that seems to be associated with amyotrophic lateral sclerosis in two different GWAS studies. _x000D_ Insight gained in this study could therefore contribute to a better understanding of the role of polymorphic inversions in the regulation of gene expression and the consequences of this understudied type of genetic variants in humans.

Date of Award	31 Oct 2014
Original language	English
Supervisor	Mario Caceres Aguilar (Director)