VDJMiner

VDJMiner analyzes immune-repertoire sequencing data to identify underlying medical conditions, predict COVID-19 severity and treatment response, and interpret T-cell receptor V(D)J gene-segment associations.

Key Features:

• Identification of underlying medical conditions: Uses immune repertoire data from COVID-19 patients to identify cancers, chronic kidney disease, autoimmune diseases, diabetes, congestive heart failure, coronary artery disease, asthma, and chronic obstructive pulmonary disease.
• Prognosis prediction: Predicts COVID-19 severity with an area under the receiver operating characteristic curve (AUC) of 0.922 for severe disease.
• Treatment response prediction: Predicts response to tocilizumab with an accuracy of 0.857 in a leave-one-out test.
• Interpretable V(D)J analysis: Interprets and scores T-cell receptor V(D)J gene segments associated with disease to reveal associations with COVID-19 severity and comorbidities.

Scientific Applications:

• Personalized risk stratification: Stratifies COVID-19 patients by severity and comorbidity profiles using immune-repertoire signatures.
• Treatment decision support: Informs assessment of likely response to therapies such as tocilizumab using immune-repertoire–derived predictors.

Methodology:

VDJMiner mines immune repertoire data from a cohort of over 1,400 COVID-19 patients and applies machine learning to sequencing-derived T-cell receptor datasets to identify patterns that correlate with medical conditions and treatment responses, with model interpretability for associational insight.