G-Bean: an ontology-graph based web tool for biomedical literature retrieval

Abstract

BackgroundCurrently, most people use NCBI's PubMed to search the MEDLINE database, an important bibliographical information source for life science and biomedical information. However, PubMed has some drawbacks that make it difficult to find relevant publications pertaining to users' individual intentions, especially for non-expert users. To ameliorate the disadvantages of PubMed, we developed G-Bean, a graph based biomedical search engine, to search biomedical articles in MEDLINE database more efficiently.MethodsG-Bean addresses PubMed's limitations with three innovations: (1) Parallel document index creation: a multithreaded index creation strategy is employed to generate the document index for G-Bean in parallel; (2) Ontology-graph based query expansion: an ontology graph is constructed by merging four major UMLS (Version 2013AA) vocabularies, MeSH, SNOMEDCT, CSP and AOD, to cover all concepts in National Library of Medicine (NLM) database; a Personalized PageRank algorithm is used to compute concept relevance in this ontology graph and the Term Frequency - Inverse Document Frequency (TF-IDF) weighting scheme is used to re-rank the concepts. The top 500 ranked concepts are selected for expanding the initial query to retrieve more accurate and relevant information; (3) Retrieval and re-ranking of documents based on user's search intention: after the user selects any article from the existing search results, G-Bean analyzes user's selections to determine his/her true search intention and then uses more relevant and more specific terms to retrieve additional related articles. The new articles are presented to the user in the order of their relevance to the already selected articles.ResultsPerformance evaluation with 106 OHSUMED benchmark queries shows that G-Bean returns more relevant results than PubMed does when using these queries to search the MEDLINE database. PubMed could not even return any search result for some OHSUMED queries because it failed to form the appropriate Boolean query statement automatically from the natural language query strings. G-Bean is available at http://bioinformatics.clemson.edu/G-Bean/index.php.ConclusionsG-Bean addresses PubMed's limitations with ontology-graph based query expansion, automatic document indexing, and user search intention discovery. It shows significant advantages in finding relevant articles from the MEDLINE database to meet the information need of the user.

A genetic algorithm-based weighted ensemble method for predicting transposon-derived piRNAs

“…Here, three latest methods: piRNApredictor [20], Piano [21] and our previous work [22] are adopted as the benchmark methods, for they build prediction models based on machine learning methods. piRNApredictor used k -mer feature (i.e, spectrum profile), k  = 1, 2, 3, 4, 5, and Piano used the LSSTE feature.…”

“…Local structure-sequence triplet elements (LSSTE): LSSTE adopts the piRNA-transposon interaction information to extract 32 different triplet elements, which contain the structural information of transposon-piRNA alignment as well as the piRNA sequence information [21, 41, 42]. …”

“…Zhang et al utilized a k -mer feature, and adopted support vector machine (SVM) to build the classifier (named piRNApredictor) for piRNA prediction [20]. Wang et al proposed a method named Piano to predict piRNAs, by using piRNA-transposon interaction information and SVM [21]. These methods exploited different features of piRNAs, and build the prediction models by using machine learning methods.…”

G-Bean: An ontology-graph based web tool for biomedical literature retrieval

“…The 106 queries consist of patient information (a brief statement about the patient) and information need (a clinician's information request statement for the patient) fields [33].…”

PSE-HMM: genome-wide CNV detection from NGS data using an HMM with Position-Specific Emission probabilities

“…HMM is also applied for modeling NGS read count data [8, 15]. In [8], an HMM with a Poisson emission probability is applied for modeling the observed read counts per genomic segment, after taking the genome-wide variation in GC contents into account.…”