Year: 2024
Type: article
Abstract: Predicting phenotypes from genotypes is a central challenge in systems biology in understanding how organisms respond to environmental and genetic changes. To address this, genome-scale metabolic models (GSMs) integrated with omics data and machine learning provide a comprehensive framework to connect genotype to phenotype. This dissertation utilizes these approaches in plants and bacterial systems, revealing key metabolic adaptations. In maize, a root-specific GSM reveals metabolic reprogramming under nitrogen stress, offering strategies for improving stress tolerance. Building on this, a multi-organ maize GSM of temperature stress identifies metabolic bottlenecks, providing targets to enhance crop resilience. Similarly, in rice, a grain-specific GSM uncovers metabolic markers tied to grain chalkiness under rising night temperatures, providing strategies for better crop quality. Moving to bacterial systems, a GSM of Rhodopseudomonas palustris explores regulatory pathways controlling cellular redox balance, while a Chlamydia trachomatis model identifies enzymes vital for persistence during nutrient deprivation. To complement these findings, machine learning models, based on omics data, predict bacterial growth responses to lignin degradation in R. palustris, identifying genes critical for growth. Together, these integrated approaches demonstrate the power of GSMs and machine learning in predicting phenotypes. Advisor: Rajib Saha
Source: Lincoln (University of Nebraska)
Author: Chowdhury, Niaz Bahar
Institutions: University of Nebraska–Lincoln, University of Nebraska System
Language: English
Subfield: Molecular Biology
Domain: Life Sciences
SDG: Zero hunger
Open Access status: green
OpenAlex ID: W7006250370
FWCI:
Cites:
Cited by:
Related to: