Using literature-based discovery to identify candidate genes for the interaction between myocardial infarction and depression

Abstract

Background A multidirectional relationship has been demonstrated between myocardial infarction (MI) and depression. However, the causal genetic factors and molecular mechanisms underlying this interaction remain unclear. The main purpose of this study was to identify potential candidate genes for the interaction between the two diseases. Methods Using a bioinformatics approach and existing gene expression data in the biomedical discovery support system (BITOLA), we defined the starting concept X as “Myocardial Infarction” and end concept Z as “Major Depressive Disorder” or “Depressive disorder”. All intermediate concepts relevant to the “Gene or Gene Product” for MI and depression were searched. Gene expression data and tissue-specific expression of potential candidate genes were evaluated using the Human eFP (electronic Fluorescent Pictograph) Browser, and intermediate concepts were filtered by manual inspection. Results Our analysis identified 128 genes common to both the “MI” and “depression” text mining concepts. Twenty-three of the 128 genes were selected as intermediates for this study, 9 of which passed the manual filtering step. Among the 9 genes, LCAT , CD4 , SERPINA1 , IL6 , and PPBP failed to pass the follow-up filter in the Human eFP Browser, due to their low levels in the heart tissue. Finally, four genes ( GNB3 , CNR1 , MTHFR , and NCAM1 ) remained. Conclusions GNB3 , CNR1 , MTHFR , and NCAM1 are putative new candidate genes that may influence the interactions between MI and depression, and may represent potential targets for therapeutic intervention.

Literature-based discovery approaches for evidence-based healthcare: a systematic review

“…Thus, Chen et al [ 37 ] have used an open LBD approach for the detection of associations among complex diseases. Closed LBD approach was also used for the explanation of the correlation between epilepsy and inflammatory bowel disease [ 38 ], and between myocardial infarction and depression [ 39 ]. Rather et al [ 40 ] proposed the use of deep learning for the discovery of potential new biomedical knowledge .…”

“…The ABC model of LBD is a common relation extraction technique used by many authors [ 18 – 20 , 26 , 30 , 31 , 39 ]. The associations between the different concepts are usually deduced from semantic predications extracted from NLP tools, like SemRep and MetaMap, which have been the most preferred tools.…”