DeSciDe: a tool for unbiased literature searching and gene list curation unveils a new role for the acidic patch mutation H2A E92K

Abstract

Omics analysis has become an indispensable tool for researchers in the life sciences, enabling the study of DNA, RNA, and proteins and how they respond to cellular stimulus. Many methods of data analysis exist for the generation and characterization of gene lists, however, selection of genes for further investigation is still heavily influenced by prior knowledge, with practitioners often studying well characterized genes, reinforcing bias in the literature. Here, we have developed an open-source, R package for impartial ranking of gene lists derived from omics analysis that we term deciphering scientific discoveries (DeSciDe). We applied a pipeline that sorts a gene list first by precedence, which we define as co-occurrence of the gene with pre-defined search terms in publications. We then rank gene lists by connectivity, an underutilized metric for how related a gene is to other enriched genes. The combination of these rankings by scatterplot provides a method for gene selection by simple visual analysis. We apply this analysis method to published Omics datasets, identifying novel avenues for investigation. Further, using this method we have been able to assign a novel loss of function role for the histone mutation H2A E92K.