Machine Learning in Computational Biology (the frequentist approach)

These lecture will introduce some general concepts and algorithms in statistical learning, illustrating them through applications in biology and personalized medicine. I will discuss linear methods in classification and regression, nonlinear extensions with positive definite kernels, and feature selection and structured sparsity. Application will include molecular diagnosis and prognosis in cancer, virtual screening in drug discovery, and biological network inference. Outline: *Introduction to pattern recognition and regression for biology and personalized medicine *Linear methods for regression and classification (OLS, RR, LDA, QDA, logistic regression, SVM...) *Nonlinear extensions with kernels *Feature selection and structured sparsity (lasso and variants) *Application: cancer prognosis from genomic data *Application: drug discovery *Application: gene networks inference