66573 - Laboratory of Bioinformatics 2

Learning outcomes

At the end of the course, the student acquires expertise on the selection and/or development and application of tools useful to address important problems of Bioinformatics and to verify the capability in handling autonomously a research project. The student will be acquainted with: - analyzing a research project where the bioinformatic approach is required; - developing the project workflow with all the necessary steps; - evaluating all the possible risks of failure and success probability; - applying of selected bioinformatics tools for the project outcomes; - developing the required software if necessary; - analyzing the results in terms of their exportability to a wet lab; - drawing conclusions in terms of benefits vs. putative costs.

Course contents

The course focuses on the development of a bioinformatics project through the application of machine learning methodologies.

The following topics will be addressed:

• Construction and curation of biological datasets (including protein sequences and/or structures)
•  Preprocessing of datasets for machine learning applications
• Protein representation techniques: one-hot encoding, sequence profiles derived from multiple sequence alignments, physicochemical feature-based representations, and advanced deep learning-based methods (e.g., embeddings)
• Manual and/or automated feature extraction and selection methodologies
• Reduction of redundancy in biological datasets, with a focus on sequence clustering techniques
• Development and implementation of predictive models in bioinformatics, including classifiers and regressors
• Strategies for hyperparameter tuning and model optimization in machine learning
• Evaluation metrics and methodologies for assessing the performance of machine learning models in bioinformatics
• Presentation and critical discussion of results

In addition, a series of lectures will cover the following foundational and advanced topics:

• Core architectures of deep neural networks (including multilayer perceptrons, convolutional neural networks, recurrent neural networks, and graph neural networks) and their applications in bioinformatics
• Attention mechanisms and transformer models, with a focus on their use in protein structure and function prediction
• Language models and protein-specific language models for high-dimensional encoding of protein sequences

Readings/Bibliography

Selected reviews and articles in cloud sharing

Teaching methods

Lectures and practicum. Development of a project in the field of Bioinformatics

Assessment methods

The final examination aims to assess the achievement of the course's learning objectives, specifically the formation of a specialist in Bioinformatics.
The candidate’s proficiency will be evaluated based on their ability to develop a bioinformatics project, to be submitted prior to the oral examination in the format of a manuscript suitable for submission to a scientific journal. The manuscript must include an Introduction, Materials and Methods, Algorithmic Development, Results, and Discussion.

The oral examination will focus on topics related to the development of the project as well as on theoretical concepts covered during the course.

Teaching tools

Online, Public Data Bases, PubMed, and materials (lecture's pdfs, selected articles)

In order to attend the course students should have taken the courses: Moduli 1 e 2 di formazione sulla sicurezza nei luoghi di studio, [https://elearning-sicurezza.unibo.it/] in e-learning.

Office hours

See the website of Castrense Savojardo