Machine Learning and Multiomics for Advanced Genomic Analysis

Speaker:  Luca Pinello - Harvard Medical School
  Thursday, December 5, 2024 at 10:30 AM

Lecture 1: Chromatin Accessibility, Multiome Assays, and Network Reconstruction  (5/12/2024  Room1.02 10:30-13:30)
This lecture will assume familiarity with scRNA-seq and dive into chromatin accessibility assays, such as ATAC-seq, which are used to identify open chromatin regions and infer gene regulation. We’ll discuss how these technologies work and introduce algorithms used for peak calling and other analyses. From there, we will explore how these methods are integrated into multiomics assays, which combine chromatin accessibility with other omics layers at the single-cell level to offer a more holistic view of cellular states. In the second part, we will briefly discuss SIMBA, a graph embedding technique our lab has developed to model relationships between single cells and multi-omic features and dictys, a method designed to reconstruct networks from multiome data, highlighting its utility in deciphering complex dynamical biological networks and regulatory interactions.

Lecture 2: Machine Learning in Biology and DNA-diffusion (6/12/2024 Room1.02 14:30-17:30)
This lecture will cover the application of machine learning and deep learning methods in biology, focusing on how they are used to analyze and interpret complex biological data. We will explore key algorithms like supervised learning and neural networks. We'll also touch on common challenges in applying these methods, such as overfitting and data bias. The second part of the lecture will introduce DNA-diffusion, a generative AI method developed in our lab to create novel regulatory DNA sequences. I will explain how diffusion models have been adapted for genomic applications and demonstrate how this method helps in designing functional regulatory elements with potential applications in synthetic biology and gene regulation studies.

 


Programme Director
Rosalba Giugno

External reference
Publication date
October 7, 2024

Studying

Share