PhD Dissertation Proposal by Akram Vasighizaker:"Manifold and representation learning for single-cell biology"

 

Date: Tuesday, April 25th, 2023

Time:  1:00pm – 3:00pm

Location: Essex Hall, Room 122

 

Reminders: 1. Two-part attendance mandatory (sign-in sheet, QR Code) 2. Arrive 5-10 minutes prior to event starting - LATECOMERS WILL NOT BE ADMITTED. Note that due to demand, if the room has reached capacity, even if you are "early" admission is not guaranteed. 3. Please be respectful of the presenter by NOT knocking on the door for admittance once the door has been closed whether the presentation has begun or not (If the room is at capacity, overflow is not permitted (ie. sitting on floors) as this is a violation of the Fire Safety code). 4. Be respectful of the decision of the advisor/host of the event if you are not given admittance. The School of Computer Science has numerous events occurring soon.

 

 

Manifold learning is a powerful technique that refers to the process of reducing and embedding high-dimensional data to lower dimensions while preserving the underlying structure of the data. In single-cell biology, high-dimensional data are typically generated from single-cell RNA sequencing (scRNA-seq) experiments, where the expression levels of thousands of genes are measured for each individual cell. Manifold learning techniques such as Modified LLE and ICA can facilitate the identification of distinct cell populations by enhancing the visualization of scRNA-seq data.

Moreover, graph representation learning is an effective technique for analyzing high-dimensional data and modelling complex interactions. Graphs can represent a variety of biological interactions, including cell-cell communication, gene regulatory networks, and protein-protein interactions. Graph neural networks (GNNs) are a popular class of models used for graph representation learning, and they have been successfully applied to single-cell biology to predict cell-cell interactions and identify rare cell types.

Combining manifold learning and graph representation learning can further improve the accuracy of cell type identification and prediction of cell-cell interactions in single-cell biology. My aim is to use nonlinear embedding primarily for single-cell RNAseq data dimensionality reduction, finding marker genes and cell types, and then improving predicting molecular interactions by combining information, including underlying cell types. It can gain deeper insights into cellular heterogeneity and the underlying mechanisms of cellular interactions which can ultimately lead to new discoveries and treatments for a variety of diseases.

 

Keywords:  manifold learning, graph representation, link prediction, single-cell biology

 

Internal Reader: Dr. Saeed Samet

Internal Reader: Dr. Dima Alhadidi

External Reader: Dr. Esam Abdel-Raheem   

Advisor: Dr. Luis Rueda