A web-based craniofacial disease gene discovery tool



 DESCRIPTION (provided by applicant): Craniofacial (CF) abnormalities constitute more than a third of all human structural birth defects. To define their genetic etiology, detailed molecular understanding is required of coordinated movement and fusion of embryonic facial prominences - as disruption of these morphogenetic events cause defects such as orofacial clefts (OFC). The NIH FaceBase initiative is an important step to address this need, as it aims to generate comprehensive whole-genome expression datasets using microarrays or Next-Gen RNA-sequencing (RNA-seq) on mouse embryonic CF tissue. However, genome-wide profiling identifies several thousand "expressed" genes and it is a formidable challenge to predict and prioritize the select few genes that are critical to tissue development or pathogenesis. We posit that although there is a wealth of genomic-level data available, this deficit remains because an adequate strategy has not yet been applied to identify these important candidate CF genes. We recently developed an innovative approach - termed in silico whole embryo body (WB) subtraction - to identify such important genes based on developmentally-enriched expression. We have applied this novel approach to ~15% of FaceBase data and assembled this knowledge as a user-friendly web-based interactive tool SysFACE (Systems tool for craniofacial expression-based gene discovery, http://bioinformatics.udel.edu/Research/SysFACE). Even with limited datasets, the beta version of SysFACE is significantly more effective, compared with unprocessed FaceBase datasets, in identification of known genes associated with OFCs from both linkage and GWAS studies. To process all existing FaceBase datasets, we will generate additional platform-specific WB reference datasets and evaluate these further with machine learning strategies to identify genes important to CF development (Aim 1). Subsequently, we aim to experimentally validate these tissue-enriched gene expression profiles, and to assemble this knowledge - along with a new evidence-based functional gene regulatory network (GRN) that will allow all molecular data from the CF published literature to be represented on systems level - as a user-friendly web-based interactive resource (Aim 2), which will also be made available through FaceBase. Development of SysFACE, as outlined in this application, will greatly improve prediction of candidate CF genes, provide an excellent resource for CF-network construction, and will facilitate CF gene discovery efforts by developmental biologists and clinicians.