LeMeDISCO is a computational method for large-scale prediction & molecular interpretation of disease comorbidity

Abstract

To understand the origin of disease comorbidity and to identify the essential proteins and pathways underlying comorbid diseases, we developed LeMeDISCO (Large-Scale Molecular Interpretation of Disease Comorbidity), an algorithm that predicts disease comorbidities from shared mode of action proteins predicted by the artificial intelligence-based MEDICASCY algorithm. LeMeDISCO was applied to predict the occurrence of comorbid diseases for 3608 distinct diseases. Benchmarking shows that LeMeDISCO has much better comorbidity recall than the two molecular methods XD-score (44.5% vs. 6.4%) and the SAB score (68.6% vs. 8.0%). Its performance is somewhat comparable to the phenotype method-based Symptom Similarity Score, 63.7% vs. 100%, but LeMeDISCO works for far more cases and its large comorbidity recall is attributed to shared proteins that can help provide an understanding of the molecular mechanism(s) underlying disease comorbidity. The LeMeDISCO web server is available for academic users at: http://sites.gatech.edu/cssb/LeMeDISCO.

PHEVIR: an artificial intelligence algorithm that predicts the molecular role of pathogens in complex human diseases

“…The precision and recall rate of LeMeDISCO co-morbidity prediction on a large set of clinical observation data (~ 200,000 pairs of diseases) are 77.2 and 37.1%, respectively. On a variety of consensus datasets, in comparison to other molecular methods 24 , 25 , LeMeDISCO has an order of magnitude larger recall rate with similar precision 21 . For pathogen-cancer associated (either oncogenic or oncolytic) virus prediction, on a set of 13 viruses including 9 known oncogenic viruses, the recall rate is 66.7% with a precision of 100% 26 .…”

“…The PHEVIR algorithm works as follows: Previously we employed the LeMeDISCO 21 algorithm to predict disease co-morbidities based on a common set of mode of action (MOA) proteins. We assert that if a viral or bacterial protein interacts with these MOA proteins, it helps cause the corresponding comorbid diseases.…”

“…Comorbidities were predicted by LeMeDISCO; we refer the reader to Ref. 21 for details. Following determination of the significant comorbidities associated with the 312 viruses and 57 bacteria curated from HPIDB, a virus-disease network and a bacteria-disease network were constructed.…”

“…To enhance the understanding of the mode of action (MOA) proteins driving the down-stream consequences associated with pathogen infection, we have developed the PHEVIR algorithm: disease comorbidities Predicted using Human–pathogEn interactomes for VIRulence. Here, we employ LeMeDISCO 21 , a recently developed tool that predicts on a proteomic scale human disease comorbidities, comorbidity enriched human MOA proteins and pathways given a pathogenic gene-human interactome set. At present, the pathogen-human interactome is provided by the HPIDB database 22 but in practice any set of human–pathogen protein–protein interactions may be used.…”