Grant: Discovery Grants Program - Individual

The cell is a complex system, orchestrated by thousands of molecular components. In higher eukaryotic cells, alternative splicing and isoform expression is a crucial layer that expands transcriptome complexity and yields functional diversity. Through alternative splicing, multiple isoforms can be generated from a single gene. It is well established that different isoforms that originated from a single gene can have unique, sometimes even opposite, functions. The advent of high-throughput sequencing techniques has allowed us to identify and quantify isoforms at the genome level. However, the precise function of most isoforms remains unknown. My research program aims to elucidate how isoforms control and drive cellular processes. To achieve this goal, we aim to develop computational and bioinformatics tools that will enable us to predict isoform functions from their sequence and expression information. Our research program will utilize novel machine learning and statistical approaches that integrate omics data to predict isoform functions. Our proposal encompasses two complementary aims, seeking to improve our understanding of isoform biology. In Aim-1, we will construct a text-mining pipeline for comprehensive scientific literature mining. The primary aim here is to curate a knowledge base comprising experimentally validated information pertaining to isoform functions. Aim-2 will focus on developing a machine learning algorithm to predict isoform functions. Our algorithm will learn from multiple isoform-related data types, such as sequence and tissue-specific isoform expression profiles. Ultimately, our algorithm will integrate the output of multiple machine learning models using an ensemble learning approach to provide an isoform-function score. Subsequently, we will implement our database and algorithm in the form of a web tool, enabling fellow researchers to seamlessly predict the function of their isoforms of interest. The outcomes of our research program will yield novel computational tools, algorithms, and knowledge resources that are instrumental in advancing our understanding of isoform biology. It will advance our understanding of molecular machinery and open novel avenues to control cellular phenotypes by targeting isoforms. Importantly, highly qualified personnel trained in this research program will acquire the most in-demand skills in machine learning, bioinformatics, and data analysis. They will gain valuable research experience and skills, preparing them for careers in both industry and academia.