📚 Syllabus & Course Curriculum
Bioinformatics & Computational Biology
Module-by-module breakdown of Cancer Risk Prediction with Machine Learning for Bioinformatics, from foundations to a certified capstone project.
Framing
• Absolute against relative risk and what a clinician can act on
• Screening, surveillance and prevention as different decision contexts
• Existing models such as Gail, Tyrer-Cuzick and BOADICEA as the baseline to beat
Data
• Germline variants, polygenic risk scores, clinical and lifestyle variables
• Ancestry bias in GWAS-derived scores and the poor transfer across populations
• Case-control sampling and how it distorts an apparent risk estimate
Modelling
• Survival models against binary classification, and the censoring that decides it
• Regularised regression and gradient boosting on tabular clinical data
• Leakage through follow-up variables that encode the outcome
Evaluation
• Discrimination, calibration and the recalibration a transferred model needs
• Decision curve analysis against current screening guidelines
• External validation in an independent cohort as the minimum credible evidence
Ethics
• Incidental findings, genetic counselling and the duty to inform
• Insurance, discrimination and the legal position in different jurisdictions
• Overdiagnosis and the harm a well-calibrated model can still cause
Earn government-registered certification in Cancer Risk Prediction with Machine Learning for Bioinformatics
e-Certificate and e-Marksheet issued on successful completion.
