An explainable machine learning data analytics method using TIGIT-linked genes for identifying biomarker signatures to clinical outcomes DOI Open Access

G Soorya,

Divya Agrawal, Shilpa Bhat

и другие.

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2023, Номер unknown

Опубликована: Дек. 7, 2023

ABSTRACT In the last decade, immunotherapies targeting immune checkpoint inhibitors have been extremely effective in eliminating subsets of some cancers patients. Multi-modal and non-immune factors that contribute to clinical outcomes utilized for predicting response therapies developing diagnostics. However, these data analytic methods involve a combination complex mathematical analytics, even-more biological mechanistic pathways. order develop method analytics transcriptomics sets, we an explainable machine learning (ML) model investigate genes involved signaling pathway T-cell-immunoreceptor with immunoglobulin ITIM domain (TIGIT). TIGIT is receptor on T, NK, T-regulatory cells, has classified as inhibitor due its ability inhibit innate adaptive responses. We extracted gene whole genome sequencing 1029 early breast cancer patient tumors, adjacent normal tissues from TCGA UCSC Xena Data Hub public databases. followed workflow which following steps: i) acquisition, processing, visualization by ii) developed predictive prognostic using input (gene expression data) output (survival time) parameters iii) interpretation was performed calculating SHAP (Shapely-Additive-exPlanations); iv) application Cox-regression model, trained L-2 regularization optimization 5 fold cross validation. The identified signatures associated predicted survival outcome test set score 0.601. summary, this case study TIGIT-mediated pathways roadmap biologists harness ML effectively.

Язык: Английский

An explainable machine learning data analytics method using TIGIT-linked genes for identifying biomarker signatures to clinical outcomes DOI Open Access

G Soorya,

Divya Agrawal, Shilpa Bhat

и другие.

bioRxiv (Cold Spring Harbor Laboratory), Год журнала: 2023, Номер unknown

Опубликована: Дек. 7, 2023

ABSTRACT In the last decade, immunotherapies targeting immune checkpoint inhibitors have been extremely effective in eliminating subsets of some cancers patients. Multi-modal and non-immune factors that contribute to clinical outcomes utilized for predicting response therapies developing diagnostics. However, these data analytic methods involve a combination complex mathematical analytics, even-more biological mechanistic pathways. order develop method analytics transcriptomics sets, we an explainable machine learning (ML) model investigate genes involved signaling pathway T-cell-immunoreceptor with immunoglobulin ITIM domain (TIGIT). TIGIT is receptor on T, NK, T-regulatory cells, has classified as inhibitor due its ability inhibit innate adaptive responses. We extracted gene whole genome sequencing 1029 early breast cancer patient tumors, adjacent normal tissues from TCGA UCSC Xena Data Hub public databases. followed workflow which following steps: i) acquisition, processing, visualization by ii) developed predictive prognostic using input (gene expression data) output (survival time) parameters iii) interpretation was performed calculating SHAP (Shapely-Additive-exPlanations); iv) application Cox-regression model, trained L-2 regularization optimization 5 fold cross validation. The identified signatures associated predicted survival outcome test set score 0.601. summary, this case study TIGIT-mediated pathways roadmap biologists harness ML effectively.

Язык: Английский

Процитировано

0