PhenoApt leverages clinical expertise to prioritize candidate genes via machine learning

Chen, Zefu; Zheng, Yu; Yang, Yongxin; Huang, Yingzhao; Zhao, Sen; Zhao, Hengqiang; Yu, Chenxi; Dong, Xian‐Zhe; Zhang, Yuanqiang; Wang, Lianlei; Zhao, Ziran; Wang, Shengru; Yang, Yang; Ming, Yue; Su, Jianzhong; Qiu, Guixing; Wu, Zhihong; Zhang, Terry Jianguo; Wu, Nan

doi:10.1016/j.ajhg.2021.12.008

Cited by 9 publications

(14 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The combined system, which we called PhenoGenius provided the best performance and reached a nearly full match (99.8%, 1682/1686) for all clinical descriptions. We compared PhenoGenius to four recently published algorithms for phenotype-driven gene prioritization: PhenoApt, Phen2Gene, CADA, and LIRICAL (7,(14)(15)(16). Despite using different prioritization methodologies, these four programs demonstrated similar performances in phenotype matching (Figure 5C).…”

Section: Conditional Algorithm Of Symptom Interaction Modeling For Ph...mentioning

confidence: 99%

“…This work has been partially supported by MIAI@Grenoble Alpes (ANR-19-P3IA-0003). 6 , Nicolas Chatron 7 , Cedric Le Maréchal 8 , Jean-François Taly 9 , Wilfrid Carre 10 , Claire Bardel 11 , Frederic Tran Mau-Them 6 , Marc Planes 12 , Marie-Pierre Audrezet 12 , Laure Raymond 9 , Charles Coutton 13 , Pierre Ray 13 , Veronique Satre 13 , Klaus Dietrich 13 , Isabelle Marey 13 , Françoise Devillard 13 , Radu Harbuz 13 , Florence Amblard 13 , Pauline Le Tanno 13 , Mouna Barat-Houari 14 , Marjolaine Willems 14 , Thomas Guignard 14 , Sylvie Odent 15 , Marie de Tayrac 10 , Damien Sanlaville 7 , Laurence Faivre 6 , Laurent Mesnard 16…”

Section: Acknowledgmentsmentioning

confidence: 99%

See 1 more Smart Citation

Learning phenotypic patterns in genetic diseases by symptom interaction modeling

Yauy

Duforet-Frebourg

Testard

et al. 2022

Preprint

View full text Add to dashboard Cite

Precision medicine relies on recognizing clinically relevant groups of patients but is currently weakened by the lack of consistency in phenotyping. Inspired by medical inductive reasoning, we developed a machine-learning system which models symptom interactions in genetic diseases to overcome this phenotyping heterogeneity. Based on the Human Phenotype Ontology and using medical data from raw text, we determine 1,1 million associated pairs of symptoms in diseases. From 16,600 phenotypic traits, we defined 390 groups of interacting symptoms summarizing the overall clinical spectrum. The symptom interaction model allowed to cover 99.8% of given symptom-gene relationships from a cohort of 1,686 diagnosed patients. This method should enable new discoveries in precision medicine by standardizing clinical descriptions.

show abstract

Section: Conditional Algorithm Of Symptom Interaction Modeling For Ph...mentioning

confidence: 99%

Section: Acknowledgmentsmentioning

confidence: 99%

Learning phenotypic patterns in genetic diseases by symptom interaction modeling

Yauy

Duforet-Frebourg

Testard

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…The cases came from five types of resources [34][35][36][37][38], such as the American Journal of Human Genetics Article, Columbia University Medical Center, and the Department of Genomic Diagnostics at the Children's Hospital of Philadelphia. For Cohorts 3-5, we used the same data cohort as Chen et al [30] group [42][43][44]. For Cohort 6, we used the same data cohort as Yuan et al [45].…”

Section: Evaluation Datasets Of True Diagnostically Validated Patientsmentioning

confidence: 99%

“…The second category of phenotype methods also aims to find relationships between phenotypes and candidate genes for prioritizing candidate genes, but without identifying the disease first. Typical examples of the second category include Phenolyzer [27], Phen2Gene [28], AMELIE [29], and PhenoApt [30]. Compared with these methods, the first category of approaches provides good interpretability, which is the focus of our work.…”

Section: Introductionmentioning

confidence: 99%

Phen2Disease: A Phenotype-driven Semantic Similarity-based Integrated Model for Disease and Gene Prioritization

Zhai

Huang

Shen

et al. 2022

Preprint

View full text Add to dashboard Cite

By utilizing the Human Phenotype Ontology (HPO), recent approaches to prioritizing disease-causing genes for patients become popular. However, these approaches do not comprehensively use information about phenotypes of diseases and patients. We present a new method called Phen2Disease that calculates similarity scores between two phenotype sets of patients and diseases by which to prioritize diseases and genes. Specifically, we calculate three scores of information content-based similarities using the phenotypes, and their combination as the respective benchmarks, and integrate them as a final score. Comprehensive experiments were conducted on six real data cohorts with 2051 cases and two simulated data cohorts with 1000 cases. Compared with the three state-of-the-art methods, if we only use phenotype information and HPO knowledge base, Phen2Disease outperformed all of them, particularly in cohorts with the less average numbers of HPO terms. We have found that patients with higher information content scores had more specific information so their predictions would be more accurate. In addition, Phen2Disease has high interpretability with ranked diseases and patient HPO terms provided.

show abstract

“…We then compared PhenoGenius to four recently published algorithms for phenotype-driven gene prioritization: PhenoApt, Phen2Gene, CADA, and LIRICAL 7,[15][16][17] . Despite using different prioritization methodologies, these four programs demonstrated similar performances in phenotype matching (Figure 5C).…”

Section: Symptom Interaction Models As An Efficient and Robust System...mentioning

confidence: 99%

Learning phenotypic patterns in genetic diseases by symptom interaction modeling

Thévenon

Yauy

Duforet-Frebourg

et al. 2022

Preprint

View full text Add to dashboard Cite

Observing phenotyping practices from an international cohort of 1,686 cases revealed heterogeneity of phenotype reporting among clinicians. Heterogeneity limited their exploitation for diagnosis as only 43% of symptom-gene associations in the cohort were available in public databases. We developed a symptom interaction model that summarized 16,600 terms into 390 groups of interacting symptoms and detected 3,222,053 novel symptom-gene associations. By learning phenotypic patterns in genetic diseases, symptom interaction modeling handled heterogeneity in phenotyping, to the extent of covering 99.8% of our cohort’s symptom-gene associations. Using these symptom interactions improved the diagnostic performance in gene prioritization by 42% (median rank 80 to 41) compared to the best algorithms. Symptom interaction modeling will provide new discoveries in precision medicine by standardizing clinical descriptions.

show abstract

PhenoApt leverages clinical expertise to prioritize candidate genes via machine learning

Cited by 9 publications

References 37 publications

Learning phenotypic patterns in genetic diseases by symptom interaction modeling

Learning phenotypic patterns in genetic diseases by symptom interaction modeling

Phen2Disease: A Phenotype-driven Semantic Similarity-based Integrated Model for Disease and Gene Prioritization

Learning phenotypic patterns in genetic diseases by symptom interaction modeling

Contact Info

Product

Resources

About