DeSciDe: Unbiased Gene Analysis Tool Uncovers H2A E92K's Role in Cell Cycle Regulation

Omics analyses, which survey changes in RNA or protein expression, generate vast datasets requiring robust filtering. However, current methods for selecting genes for further investigation are often influenced by prior knowledge, leading researchers to frequently study well-characterized genes. This perpetuates a significant bias in the literature, leading to many potentially crucial genes remaining underexplored. Recognizing this urgent need for unbiased tools, researchers have developed DeSciDe (deciphering scientific discoveries), an open-source R package aimed at providing impartial ranking for gene lists derived from omics analyses.

DeSciDe addresses the challenge of bias by incorporating both literature precedence and network connectivity into its ranking system. The tool first sorts genes by 'precedence,' defined by their co-occurrence with user-defined search terms in PubMed abstracts, allowing for context-specific filtering. Subsequently, genes are ranked by 'connectivity,' an underutilized metric quantifying how related a gene is to other enriched genes within interaction networks like those from STRING. By combining these rankings into an interactive scatterplot, DeSciDe offers a simple visual method for gene selection, allowing researchers to quickly identify both highly studied and underexplored, yet highly connected, genes relevant to their cellular stimulus of interest.

Applying DeSciDe to published omics datasets has yielded significant insights, notably revealing a novel loss-of-function role for the histone mutation H2A E92K. Through unbiased analysis of a proximity proteomics dataset, DeSciDe's pipeline indicated that this acidic patch mutation leads to G2 phase stalling and a subsequent reduction in cell proliferation. This discovery exemplifies the tool's utility, as lead author Douglas notes in the paper, "This example indicates that unbiased analysis using a program such as DeSciDe can provide different avenues of investigation that may be overlooked in favour of genes that are highly represented in the literature."