Two Similarity Metrics for Medical Subject Headings (MeSH): An Aid to Biomedical Text Mining and Author Name Disambiguation

Smalheiser, Neil R.; Bonifield, Gary

doi:10.1101/039008

Cited by 4 publications

(4 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To characterize these articles and identify any differences, we extracted other MeSH terms associated with each of the two corpora. Smalheiser and Bonifield recently proposed a metric for quantifying the semantic relatedness of two MeSH terms through their tendency to co‐occur in the same article's MEDLINE entry 21 . We used this metric and ranked associated MeSH terms according to their odd ratio and kept the top 40 associated MeSH terms (Table ).…”

Section: Methodsmentioning

confidence: 99%

Atopic dermatitis or eczema? Consequences of ambiguity in disease name for biomedical literature mining

Frainay

Pitarch

Filippi

et al. 2021

Clin Experimental Allergy

View full text Add to dashboard Cite

Background Biomedical research increasingly relies on computational approaches to extract relevant information from large corpora of publications. Objective To investigate the consequence of the ambiguity between the use of terms “Eczema” and “Atopic Dermatitis” (AD) from the Information Retrieval perspective, and its impact on meta‐analyses, systematic reviews and text mining. Methods Articles were retrieved by querying the PubMed using terms ‘eczema’ (D003876) and “dermatitis, atopic” (D004485). We used machine learning to investigate the differences between the contexts in which each term is used. We used a decision tree approach and trained model to predict if an article would be indexed with eczema or AD tags. We used text‐mining tools to extract biological entities associated with eczema and AD, and investigated the discrepancy regarding the retrieval of key findings according to the terminology used. Results Atopic dermatitis query yielded more articles related to veterinary science, biochemistry, cellular and molecular biology; the eczema query linked to public health, infectious disease and respiratory system. Medical Subject Headings terms associated with “AD” or “Eczema” differed, with an agreement between the top 40 lists of 52%. The presence of terms related to cellular mechanisms, especially allergies and inflammation, characterized AD literature. The metabolites mentioned more frequently than expected in articles with AD tag differed from those indexed with eczema. Fewer enriched genes were retrieved when using eczema compared to AD query. Conclusions and Clinical Relevance There is a considerable discrepancy when using text mining to extract bio‐entities related to eczema or AD. Our results suggest that any systematic approach (particularly when looking for metabolites or genes related to the condition) should be performed using both terms jointly. We propose to use decision tree learning as a tool to spot and characterize ambiguity, and provide the source code for disambiguation at https://github.com/cfrainay/ResearchCodeBase.

show abstract

Section: Methodsmentioning

confidence: 99%

Atopic dermatitis or eczema? Consequences of ambiguity in disease name for biomedical literature mining

Frainay

Pitarch

Filippi

et al. 2021

Clin Experimental Allergy

View full text Add to dashboard Cite

show abstract

“…Apart from author name disambiguation, what other types of text modeling tasks might benefit from using our implicit term and text similarity metrics? Word2vec-based metrics have been very actively explored for a variety of extrinsic tasks such as named entity recognition, part of speech tagging, ranking PubMed articles for semantic relatedness [18,26], and word sense disambiguation [37]. One should be cautioned that performance on one set of tasks does not necessarily correlate with performance on other tasks [38,39], yet we feel that our implicit term similarity metrics should be explored to see if they have any complementary value in modeling the latter applications.…”

Section: Discussionmentioning

confidence: 99%

“…For example, recently we created journal similarity metrics that relate any two journals A and B according to a) how similar they are topically, b) how similar they are in terms of sharing the same authors, and c) how likely it is that an author who publishes in A will publish later in B [17]. We also created MeSH (Medical Subject Headings) similarity metrics that relate any two MeSH terms A and B according to a) how often the two terms co-occur in a single article, and b) how often the two terms are found within the body of articles published by a single author [18]. Such implicit metrics are valuable in text mining models such as are used in the Author-ity author name disambiguation project [19,20], where the goal is to consider two articles that share the same author [lastname, first initial] and estimate the probability that the two articles refer to the same author-individual.…”

Section: Introductionmentioning

confidence: 99%

Unsupervised low-dimensional vector representations for words, phrases and text that are transparent, scalable, and produce similarity metrics that are not redundant with neural embeddings

Smalheiser

Cohen

Bonifield

2019

Journal of Biomedical Informatics

Self Cite

View full text Add to dashboard Cite

Neural embeddings are a popular set of methods for representing words, phrases or text as a low dimensional vector (typically 50-500 dimensions). However, it is difficult to interpret these dimensions in a meaningful manner, and creating neural embeddings requires extensive training and tuning of multiple parameters and hyperparameters. We present here a simple unsupervised method for representing words, phrases or text as a low dimensional vector, in which the meaning and relative importance of dimensions is transparent to inspection. We have created a near-comprehensive vector representation of words, and selected bigrams, trigrams and abbreviations, using the set of titles and abstracts in PubMed as a corpus. This vector is used to create several novel implicit word-word and text-text similarity metrics. The implicit word-word similarity metrics correlate well with human judgement of word pair similarity and relatedness, and outperform or equal all other reported methods on a variety of biomedical benchmarks, including several implementations of neural embeddings trained on PubMed corpora. Our implicit word-word metrics capture different aspects of word-word relatedness than word2vecbased metrics and are only partially correlated (rho = ~0.5-0.8 depending on task and corpus).The vector representations of words, bigrams, trigrams, abbreviations, and PubMed title+abstracts are all publicly available from http://arrowsmith.psych.uic.edu for release under CC-BY-NC license. Several public web query interfaces are also available at the same site, including one which allows the user to specify a given word and view its most closely related terms according to direct co-occurrence as well as different implicit similarity metrics.

show abstract

“…In order to identify drug mechanisms supported by colocalizations and to have a quantitative assessment of tissue-trait similarity, we used similarity in the MeSH vocabulary. Similarity is determined through odds ratio of cooccurance in article MeSH terms in the PubMed corpus (accessed October 10, 2018), an approach based on [13]. For identifying similar drug mechanisms, we used an odds ratio cutoff of 20 as this corresponded on average to the similarity cutoff used in the previous work.…”

Section: Pubmed Odds Ratio Similaritymentioning

confidence: 99%

Estimating colocalization probability from limited summary statistics

King

Dunbar

Davis

et al. 2020

Preprint

View full text Add to dashboard Cite

A common approach to understanding the mechanisms of noncoding GWAS associations is to test the GWAS variant for association with lower level cellular phenotypes such as gene expression. However, significant association to gene expression will often arise from linkage disequilibrium to a separate causal variant and be unrelated to the mechanism underlying the GWAS association. Colocalization is a statistical genetic method used to determine whether the same variant is causal for multiple phenotypes and is stronger evidence for understanding mechanism than shared significance. Current colocalization methods require full summary statistics for both traits, limiting their use with the majority of reported GWAS associations (e.g. GWAS Catalog). We propose a new approximation to the popular coloc method [1] that can be applied when limited summary statistics are available, as in the common scenario where a GWAS catalog hit would be tested for colocalization with a GTEx eQTL. Our method (POint EstiMation of Colocalization -POEMColoc) imputes missing summary statistics using LD structure in a reference panel, and performs colocalization between the imputed statistics and full summary statistics for a second trait.

show abstract

Two Similarity Metrics for Medical Subject Headings (MeSH): An Aid to Biomedical Text Mining and Author Name Disambiguation

Abstract: In the present paper, we have created and characterized several similarity metrics for relating any

Cited by 4 publications

References 15 publications

Atopic dermatitis or eczema? Consequences of ambiguity in disease name for biomedical literature mining

Atopic dermatitis or eczema? Consequences of ambiguity in disease name for biomedical literature mining

Unsupervised low-dimensional vector representations for words, phrases and text that are transparent, scalable, and produce similarity metrics that are not redundant with neural embeddings

Estimating colocalization probability from limited summary statistics

Contact Info

Product

Resources

About