242: Gene Set Enrichment Investigation of Maternal Exome Variation in Spontaneous Preterm Birth (SPTB)

Tracy A. Manuck,Scott Watkins,M. Sean Esplin,Samuel Parry,Heping Zhang,Hao Huang,Joseph R. Biggio,Radek Bukowski,George Saade,William Andrews,Don Baldwin,Yoel Sadovsky,Uma Reddy,John Ilekis,Michael W. Varner,Lynn B. Jorde,Mark Yandell
DOI: https://doi.org/10.1016/j.ajog.2015.10.280
IF: 9.8
2016-01-01
American Journal of Obstetrics and Gynecology
Abstract:As discovery science has moved from candidate gene testing to systems biology, several powerful new gene set algorithms have been developed to identify associated biologic pathways. We hypothesized that these new algorithms could identify promising biologic pathways associated with SPTB. Exome sequences from 138 women from GPN with ≥1 SPTB <37 wks (cases) were compared to 1253 females from the 1000 Genomes Project (controls) using Variant Annotation, Analysis & Search Tool (VAAST), a probabilistic search tool for identifying disease-causing variants. The median GA of the earliest SPTB was 30.1 wks. Genes scored by VAAST with p<1e-4 were then analyzed by 3 online tools: (1) Protein Analysis Through Evolutionary Relationships (PANTHER) (2) Database for Annotation, Visualization, and Integrated Discovery (DAVID), and (3) Disease Association Protein-Protein Link Evaluator (DAPPLE). These tools group genes into gene sets, pathways, & gene ontology (GO) groups. They assess for over-/under- representation (PANTHER), gene groupings (DAVID), and connectivity between known gene protein products (DAPPLE). VAAST results were also prioritized using Phenotype Driven Variant Ontological Re-Ranking Tool (PHEVOR) and evaluated for burden using GROUPER; both of these new VAAST algorithms used an established PTB candidate gene list (dbPTB). 440 genes identified by VAAST (recessive inheritance, locus heterogeneity) had p<1e-4. PANTHER revealed that more genes are categorized into GO Reproduction and Death groups than expected by chance (Table). Selected candidate PTB genes in each GO group are also shown in the Table; genes marked (*) were also significant using the PheVor and/or GROUPER algorithms. DAVID grouped 160/440 (40%) genes into 11 distinct clusters. DAPPLE revealed multiple connections between anticipated protein products and prioritized 8 new genes of interest (Figure). A systematic and integrated analysis of maternal exome variation using multiple new gene set algorithms independently identifies enrichment in several genetic pathways, implicates new potential candidate genes and confirms previously suspected SPTB genes. Results are consistent across independent gene prioritization algorithms. These findings require further confirmation in other populations.View Large Image Figure ViewerDownload Hi-res image Download (PPT)
What problem does this paper attempt to address?