Enhancing Phenotype Recognition in Clinical Notes Using Large Language Models: PhenoBCBERT and PhenoGPT

Overview

Journal ArXiv

Date 2023 Nov 21

PMID 37986722

Authors

Jingye Yang

Cong Liu

Wendy Deng

Da Wu

Chunhua Weng

Yunyun Zhou

Kai Wang

Affiliations

Soon will be listed here.

Abstract

To enhance phenotype recognition in clinical notes of genetic diseases, we developed two models - PhenoBCBERT and PhenoGPT - for expanding the vocabularies of Human Phenotype Ontology (HPO) terms. While HPO offers a standardized vocabulary for phenotypes, existing tools often fail to capture the full scope of phenotypes, due to limitations from traditional heuristic or rule-based approaches. Our models leverage large language models (LLMs) to automate the detection of phenotype terms, including those not in the current HPO. We compared these models to PhenoTagger, another HPO recognition tool, and found that our models identify a wider range of phenotype concepts, including previously uncharacterized ones. Our models also showed strong performance in case studies on biomedical literature. We evaluated the strengths and weaknesses of BERT-based and GPT-based models in aspects such as architecture and accuracy. Overall, our models enhance automated phenotype detection from clinical texts, improving downstream analyses on human diseases.

References

Kohler S, Schulz M, Krawitz P, Bauer S, Dolken S, Ott C . Clinical diagnostics in human genetics with semantic similarity searches in ontologies. Am J Hum Genet. 2009; 85(4):457-64. PMC: 2756558. DOI: 10.1016/j.ajhg.2009.09.003. View

Aronson A . Effective mapping of biomedical text to the UMLS Metathesaurus: the MetaMap program. Proc AMIA Symp. 2002; :17-21. PMC: 2243666. View

Anazi S, Maddirevula S, Salpietro V, Asi Y, Alsahli S, Alhashem A . Expanding the genetic heterogeneity of intellectual disability. Hum Genet. 2017; 136(11-12):1419-1429. DOI: 10.1007/s00439-017-1843-2. View

Martinez-Romero M, Jonquet C, OConnor M, Graybeal J, Pazos A, Musen M . NCBO Ontology Recommender 2.0: an enhanced approach for biomedical ontology recommendation. J Biomed Semantics. 2017; 8(1):21. PMC: 5463318. DOI: 10.1186/s13326-017-0128-y. View

Hartley T, Lemire G, Kernohan K, Howley H, Adams D, Boycott K . New Diagnostic Approaches for Undiagnosed Rare Genetic Diseases. Annu Rev Genomics Hum Genet. 2020; 21:351-372. DOI: 10.1146/annurev-genom-083118-015345. View

Sung M, Jeong M, Choi Y, Kim D, Lee J, Kang J . BERN2: an advanced neural biomedical named entity recognition and normalization tool. Bioinformatics. 2022; 38(20):4837-4839. PMC: 9563680. DOI: 10.1093/bioinformatics/btac598. View

Yang X, Chen A, PourNejatian N, Shin H, Smith K, Parisien C . A large language model for electronic health records. NPJ Digit Med. 2022; 5(1):194. PMC: 9792464. DOI: 10.1038/s41746-022-00742-2. View

Arbabi A, Adams D, Fidler S, Brudno M . Identifying Clinical Terms in Medical Text Using Ontology-Guided Machine Learning. JMIR Med Inform. 2019; 7(2):e12596. PMC: 6533869. DOI: 10.2196/12596. View

Deisseroth C, Birgmeier J, Bodle E, Kohler J, Matalon D, Nazarenko Y . ClinPhen extracts and prioritizes patient phenotypes directly from medical records to expedite genetic disease diagnosis. Genet Med. 2018; 21(7):1585-1593. PMC: 6551315. DOI: 10.1038/s41436-018-0381-1. View

10.

Peng C, Dieck S, Schmid A, Ahmad A, Knaus A, Wenzel M . CADA: phenotype-driven gene prioritization based on a case-enriched knowledge graph. NAR Genom Bioinform. 2021; 3(3):lqab078. PMC: 8415429. DOI: 10.1093/nargab/lqab078. View

11.

Soysal E, Wang J, Jiang M, Wu Y, Pakhomov S, Liu H . CLAMP - a toolkit for efficiently building customized clinical natural language processing pipelines. J Am Med Inform Assoc. 2017; 25(3):331-336. PMC: 7378877. DOI: 10.1093/jamia/ocx132. View

12.

Robinson P, Ravanmehr V, Jacobsen J, Danis D, Zhang X, Carmody L . Interpretable Clinical Genomics with a Likelihood Ratio Paradigm. Am J Hum Genet. 2020; 107(3):403-417. PMC: 7477017. DOI: 10.1016/j.ajhg.2020.06.021. View

13.

Son J, Xie G, Yuan C, Ena L, Li Z, Goldstein A . Deep Phenotyping on Electronic Health Records Facilitates Genetic Diagnosis by Clinical Exomes. Am J Hum Genet. 2018; 103(1):58-73. PMC: 6035281. DOI: 10.1016/j.ajhg.2018.05.010. View

14.

Luo L, Yan S, Lai P, Veltri D, Oler A, Xirasagar S . PhenoTagger: a hybrid method for phenotype concept recognition using human phenotype ontology. Bioinformatics. 2021; 37(13):1884-1890. PMC: 11025364. DOI: 10.1093/bioinformatics/btab019. View

15.

Groza T, Kohler S, Doelken S, Collier N, Oellrich A, Smedley D . Automatic concept recognition using the human phenotype ontology reference and test suite corpora. Database (Oxford). 2015; 2015. PMC: 4343077. DOI: 10.1093/database/bav005. View

16.

Zanello G, Chan C, Pearce D . Recommendations from the IRDiRC Working Group on methodologies to assess the impact of diagnoses and therapies on rare disease patients. Orphanet J Rare Dis. 2022; 17(1):181. PMC: 9078009. DOI: 10.1186/s13023-022-02337-2. View

17.

Zhao M, Havrilla J, Peng J, Drye M, Fecher M, Guthrie W . Development of a phenotype ontology for autism spectrum disorder by natural language processing on electronic health records. J Neurodev Disord. 2022; 14(1):32. PMC: 9128253. DOI: 10.1186/s11689-022-09442-0. View

18.

Groft S, Posada M, Taruscio D . Progress, challenges and global approaches to rare diseases. Acta Paediatr. 2021; 110(10):2711-2716. DOI: 10.1111/apa.15974. View

19.

Yang H, Robinson P, Wang K . Phenolyzer: phenotype-based prioritization of candidate genes for human diseases. Nat Methods. 2015; 12(9):841-3. PMC: 4718403. DOI: 10.1038/nmeth.3484. View

20.

Smedley D, Robinson P . Phenotype-driven strategies for exome prioritization of human Mendelian disease genes. Genome Med. 2015; 7(1):81. PMC: 4520011. DOI: 10.1186/s13073-015-0199-2. View