1. Enhancing Bacterial Genomic Analyses with Graph Genome Assemblies

This project explores the application of graph genome assemblies to improve next-generation sequencing (NGS) read alignment in bacterial genomics. Due to rapid evolutionary rates and high genomic diversity in bacteria, traditional linear reference genomes may underperform. By constructing graph genomes for representative strains of a species, the project investigates improvements in RNA-seq alignment rates and assesses how genome diversity correlates with alignment efficacy. The results will provide insights into the utility of graph genomes in bacterial species beyond E. coli, informing genomic analysis workflows in microbiology.

2. Comparative Analysis of Proteome Abundance Patterns Across Species Using Gene Ontology

This study analyzes mass spectrometry data from E. coli and yeast to explore how normalized protein abundances differ across Gene Ontology (GO) categories. Proteins are mapped to GO terms to compare biological processes and molecular functions between the two species. Non-parametric statistical tests quantify the significance of distributional differences across GO categories, providing insights into species-specific proteome strategies. The project highlights how biochemical priorities diverge even among organisms with shared molecular building blocks.

3. Containerization of Repeat-Aware Analysis Pipelines

To address limitations in existing genomic analysis pipelines that overlook transposable elements (TEs), this project enhances a repeat-aware pipeline by integrating Allo into the ENCODE ChIP-seq framework. It focuses on containerizing the pipeline using Docker and Singularity to improve reproducibility and portability. By comparing outputs across native and containerized environments, the project evaluates performance consistency while enabling broader deployment of TE-inclusive analyses, which are critical for understanding genome evolution and regulation.

4. Comparing AlphaFold2 and ESMFold Structure Predictions

This project evaluates the structural accuracy of ESMFold—a fast, alignment-free protein structure predictor—against AlphaFold2, which is widely regarded for its high accuracy. Using predictions for E. coli and yeast proteins, the study compares metrics like RMSD, native residue contacts, and residue clashes, treating AlphaFold2 as the benchmark. By assessing when speedier methods like ESMFold suffice and when AlphaFold2 remains necessary, this work informs optimal use cases for different protein modeling tools in structural biology.

5. Exploring Protein Entanglement Within AlphaFold-Predicted Structures

Focusing on protein topologies, this project quantifies native entanglements—structural motifs involving loops and threading residues—in AlphaFold-predicted proteomes. A new organism’s proteome is analyzed and compared with published entanglement profiles from E. coli, yeast, and humans. Using EntanGoPy, the project identifies entanglements, clusters structural features, and applies statistical tests to detect significant differences. Findings may illuminate evolutionary and functional consequences of entanglement motifs across species.

6. Exploring Intrinsically Disordered Proteins and Their Properties

This project investigates intrinsically disordered regions (IDRs) in the proteomes of three organisms using predictive tools like metapredict, CIDER, and ALBATROSS. IDRs lack fixed structures but play critical roles in cellular signaling and stress responses. The project quantifies IDR abundance, sequence composition, and predicted structural properties, comparing distributions across organisms. By uncovering statistically significant differences, the study provides insights into how disorder contributes to functional diversity in protein biology.