Artificial Intelligence for Bioinformatics (Blanzieri, Passerini, Tebaldi)
The 20 hours long course will be organized in three modules covering complementary aspects of AI applications to bioinformatics. The course will be also offered to the students of the transdisciplinary program in computational biology. The course does not require funding being two modules given by internal personnel and the third module included in the visiting researcher proposal by Toma Tebaldi.
1) Machine learning techniques for classification and regression tasks in bioinformatics (Enrico Blanzieri, 8 hours)
The module will cover data and nature of the tasks, local, max-margin, and neural techniques for solving them. We will also illustrate examples from SVMs for RNA-protein binding prediction to the recent applications of deep learning to protein function classification.
2) Probabilistic graphical models for bioinformatics (Andrea Passerini, 6 hours)
The basic principles underlying Bayesian Networks will be presented: representation formalism, independences, basics of inference and learning. Hands-on experience in developing and using Bayesian Networks will be provided. Finally, profile Hidden Markov Models for biological sequences will be discussed.
3) Artificial intelligence techniques for the analysis and interpretation of single cell sequencing data (Toma Tebaldi, 6 hours)
Thanks to the revolution of single cell sequencing, today we can obtain genomic and transcriptomic sequencing data from single cells. By looking at thousands of cells one at a time, we can see which set of genes each individual cell is transcribing, and we can capture the cellular diversity of human tissues with unprecedented resolution. Single cell data analysis requires the development of appropriate methods, for example for cell type identification and inference of gene regulatory networks. We will present, discuss and test some of the available techniques addressing the analysis of single cell sequencing data.