Estimating a Ranked List of Human Genetic Diseases by Associating Phenotype-Gene with Gene-Disease Bipartite Graphs

Abstract

With vast amounts of medical knowledge available on the Internet, it is becoming increasingly practical to help doctors in clinical diagnostics by suggesting plausible diseases predicted by applying data and text mining technologies. Recently, Genome-Wide Association Studies (GWAS) have proved useful as a method for exploring phenotypic associations with diseases. However, since genetic diseases are difficult to diagnose because of their low prevalence, large number, and broad diversity of symptoms, genetic disease patients are often misdiagnosed or experience long diagnostic delays. In this article, we propose a method for ranking genetic diseases for a set of clinical phenotypes. In this regard, we associate a phenotype-gene bipartite graph (PGBG) with a gene-disease bipartite graph (GDBG) by producing a phenotype-disease bipartite graph (PDBG), and we estimate the candidate weights of diseases. In our approach, all paths from a phenotype to a disease are explored by considering causative genes to assign a weight based on path frequency, and the phenotype is linked to the disease in a new PDBG. We introduce the Bidirectionally induced Importance Weight (BIW) prediction method to PDBG for approximating the weights of the edges of diseases with phenotypes by considering link information from both sides of the bipartite graph. The performance of our system is compared to that of other known related systems by estimating Normalized Discounted Cumulative Gain (NDCG), Mean Average Precision (MAP), and Kendall's tau metrics. Further experiments are conducted with well-known TF·IDF, BM25, and Jenson-Shannon divergence as baselines. The result shows that our proposed method outperforms the known related tool Phenomizer in terms of NDCG@10, NDCG@20, MAP@10, and MAP@20; however, it performs worse than Phenomizer in terms of Kendall's tau-b metric at the top-10 ranks. It also turns out that our proposed method has overall better performance than the baseline methods.

Diagnosis support systems for rare diseases: a scoping review

“…For example, when tested on simulated patients with different levels of noisy phenotype concepts, Phenomizer classified at least 75% of correct diseases as top 1, whereas when tested by another team [77], the method used in Phenomizer (best match average method) on a dataset consisting of 462 EHRs reached less than 10% of the top 10 correct disease rankings. The impact of changing the dataset for evaluation was confirmed when Phenomizer was tested for comparison with developed tools in numerous studies and obtained results that highly depended on the dataset under study [2,55,[68][69][70][71]77].…”

“…The diagnosis support system then returns a list of diseases ranked by the similarity score for each patient. Three studies [65,67,68] out of 14 included gene-disease knowledge in their model. One of these systems [67] needed as input the list of the patient's phenotype concepts complemented by the list of variants identified in the patient's genome.…”

“…Among the 11 remaining studies, 5 studies considered simulated patients (queries consisted of a list of phenotype concepts). These studies aimed to diagnose genetic [63,64,68] and rare [66,70] diseases from queries with phenotype concepts. Six studies did not provide any evaluation protocol [26,32,38,50,65,69], among which four studies developed manually designed algorithms.…”

“…Flowchart of the screening processArticles were categorized into three groups based on the number of targeted diseases:-Group 1 -Studies focusing on one disease (e.g., Huntington disease): 29 studies out of 61 (48%); -Group 2 -Studies focusing on a class of diseases or syndromes (e.g., macular diseases, facial genetic dysmorphologies): 15 studies (25%)[48][49][50][51][52][53][54][55][56][57][58][59][60][61][62]; -Group 3 -Studies focusing on the whole spectrum of genetic/rare diseases: 17 studies (28%)[2,[63][64][65][66][67][68][69][70][71][72][73][74][75][76][77][78].…”

A Bipartite Graph-Based Ranking Approach to Query Subtopics Diversification Focused on Word Embedding Features

“…(1) Bipartite Graph Based Ranking Many real applications can be modeled as a bipartite graph, including Video shots and Tags [46], Queries and URLs, Entities and Co-List [47] in a Web page, Phenotypes and Diseases [48].…”