A Survey On Deep Learning For Drug-Target Binding Prediction: Models, Benchmarks, Evaluation, And Case Studies, Kusal Debnath, Pratip Rana, Preetam Ghosh

Computer Science Faculty Publications

Conventional drug discovery is expensive, time-consuming, and prone to failure. Artificial intelligence has become a potent substitute over the last decade, providing strong answers to challenging biological issues in this field. Among these difficulties, drug-target binding (DTB) is a key component of drug discovery techniques. In this context, drug-target affinity and drug–target interaction are complementary and essential frameworks that work together to improve our comprehension of DTB dynamics. In this work, we thoroughly analyze the most recent deep learning models, popular benchmark datasets, and assessment metrics for DTB prediction. We look at the paradigm shift in the development of drug …

18s Metabarcode Analyses Of Eukaryotic Species In The Respiratory Microbiomes Of Wild Canids From New Hampshire, Collin Sinclair Blake

Honors Theses and Capstones

This study is of an exploratory nature and focuses on characterizing the eukaryotic microbiota present in the respiratory tissues of six wild canids and one domestic canine. The contents of this document largely pertain to dry lab analyses of 18S barcodes in bioinformatics programs – primarily QIIME2 running in the GitBash command line, service for which was hosted by the UNH Ron Bioinformatics training server. All procedures listed within the section below were performed by second parties at the UNH Hubbard Center for Genomics Studies (HCGS), the New Hampshire Veterinary Diagnostics Lab (NHVDL), and the UNH Microbial Ecology and Emerging …

Mobula, Bioinspiration, Filter Feeding, Form And Function, J. B. Teeple, S. R. Kahane-Rapport, K. E. Cohen, L. Hamann, J. A. Strother, E. W.M. Paig-Tran

Biological Sciences Faculty Publications

Mobulas (manta and devil rays) are large-scale ram filter feeders that separate planktonic food particles from large volumes of water with minimal clogging. This contrasts with most human-made filters that can suffer from problematic clogging requiring additional mechanisms for clearing blocked surfaces and maintaining performance. Prior studies have shown that mobulas employ a unique mechanism referred to as ricochet separation to filter feed, whereby captive vortices in filter pores cause particles to bounce off the filter surfaces and away from the filter pores. This mechanism enables the filtration of particles smaller than the pore size and reduced clogging. However, few …

Computational Tools For Protein Classification In Metagenomes, Fawad Ullah

Dissertations, Master's Theses and Master's Reports

Advances in genomic sequencing have dramatically increased the amount and the speed at which genomic data is being generated. These technological advances have enabled the ability to profile the genetic information of organisms and communities at unprecedented scales. Many methods have been developed to identify and classify genes within these datasets. However, many generic pipelines for gene annotation struggle to accurately predict specific protein classes that may not be represented in their databases. Two major challenges exist for classification of specific protein classes in metagenomic databases. The first is the fact that many metagenomic assemblies are highly fragmented with many …

Genomic Epidemiology Of Staphylococcus Aureus Sequence-Type 72, Sachitaa Senthilkumar

Honors Undergraduate Theses

Staphylococcus aureus (S. aureus, SA) is a gram-positive bacterial colonizer and pathogen commonly carried on the skin and in the anterior nares of humans. SA colonization may lead to a variety of non-invasive (e.g., skin and soft tissue infections, SSTIs) and invasive (e.g., bacteremia and osteomyelitis) infections. The SA population is comprised of multiple lineages that can be delineated through multi-locus sequence typing (MLST), which compares nucleotide sequences within seven housekeeping genes. Strains belonging to an MLST share a common evolutionary history and phenotypic characteristics like antibiotic resistance or virulence. As a result, the epidemiology of lineages can be …

Characterizing Somatic Variants In Nanopore Data With Machine Learning, Shwethal Sayeeram Trikannad

Master's Projects

Oxford Nanopore Technology (ONT) is a popular long-read sequencer in genomics. However, its high base-calling error rate produces several sequencing artifacts. Detection of somatic variants in ONT sequenced tumor-normal samples remains challenging due to low frequencies. In this study, machine learning was applied to a dataset created by benchmarking ClairS output against HCC1395 and colo829 truth sets to classify variants and artifacts. Relevant features were engineered from sequence context and variant site characteristics to model artifact profiles. HistGradientBoostingClassifier achieved 0.876950 accuracy, outperforming all other models. Variant quality was the top predictor with an aggregate accuracy of over 85%. This work …

Phylogenetic Analysis Of Metabolic Enzymes In Hypoxic/Anoxic Conditions In Cetaceans And Cancer, Samerna A. Masih

Electronic Theses & Dissertations (2024 - present)

Cancer cells often exhibit a Warburg-like metabolism, which includes increased fatty acid synthesis and storage that help them survive in hypoxic conditions. This shift is marked by the heightened synthesis and accumulation of fatty acids, acting as a survival mechanism in low-oxygen environments (Baumann et al., 2016). Interestingly, deep-diving whales may utilize a similar metabolic pathway to produce wax esters during their prolonged dives.

Our study focused on exploring the genetic and metabolic similarities between breast cancer cells and deep-diving whales, using computational analyses to investigate lipid metabolism genes. We looked for evolutionarily conserved variations that might reveal adaptive mechanisms …

Enhancing Clinical Trial Matching In Molecular Diagnostics: Using Natural Language Processing And Clustering Approaches In Hematological Malignancies, Gillian Fanning

Theses and Dissertations

Clinical trial matching is a critical component of personalized medicine, particularly in the management of hematologic malignancies. At Virginia Commonwealth University (VCU) Health, the Molecular Diagnostics (MDX) Lab produces somatic variant reports and recommends clinical trials based on the presence of clinically significant mutations. However, the current manual trial recommendation process is time-intensive and lacks scalability.

This study introduces a computational framework to streamline and standardize clinical trial matching using natural language processing (NLP) and unsupervised clustering. Trial brief descriptions were analyzed to extract frequent terms, and trials were grouped based on term similarity using joint dimensionality reduction and clustering. …

Cath-Ddg: Towards Robust Mutation Effect Prediction On Protein-Protein Interactions Out Of Cath Homologous Superfamily, Guanglei Yu, Xuehua Bi, Teng Ma, Yaohang Li, Jianxin Wang

Computer Science Faculty Publications

Motivation: Protein-protein interactions (PPIs) are fundamental aspects in understanding biological processes. Accurately predicting the effects of mutations on PPIs remains a critical requirement for drug design and disease mechanistic studies. Recently, deep learning models using protein 3D structures have become predominant for predicting mutation effects. However, significant challenges remain in practical applications, in part due to the considerable disparity in generalization capabilities between easy and hard mutations. Specifically, a hard mutation is defined as one with its maximum TM-score < 0.6 when compared to the training set. Additionally, compared to physics-based approaches, deep learning models may overestimate performance due to potential data leakage.

Results: We propose new training/test splits that mitigate data leakage according to the CATH homologous superfamily. Under the constraints of physical …

Heterogeneous Clustering Of Multiomics Data For Breast Cancer Subgroup Classification And Detection, Joseph Pateras, Musaddiq Lodi, Pratip Rana, Preetam Ghosh

Computer Science Faculty Publications

The rapid growth of diverse -omics datasets has made multiomics data integration crucial in cancer research. This study adapts the expectation–maximization routine for the joint latent variable modeling of multiomics patient profiles. By combining this approach with traditional biological feature selection methods, this study optimizes latent distribution, enabling efficient patient clustering from well-studied cancer types with reduced computational expense. The proposed optimization subroutines enhance survival analysis and improve runtime performance. This article presents a framework for distinguishing cancer subtypes and identifying potential biomarkers for breast cancer. Key insights into individual subtype expression and function were obtained through differentially expressed gene …

Deepssetracer 2.0: Improved Deep Learning Model Performance For Protein Secondary Structure Segmentation From Cryo-Em Maps, Bryan Hawickhorst, Thu Nguyen, Willy Wriggers, Jiangwen Sun, Jing He

Computer Science Faculty Publications

DeepSSETracer is a method for segmenting protein secondary structure from medium-resolution (5-10Å) cryogenic electron microscopy (cryo-EM) density maps. We conducted experiments and ablation studies to examine the effects of normalization methods, max-pooling, activation functions, and loss calculation region on DeepSSETracer. By combining multiple technical improvements, the performance of the new version, DeepSSETracer 2.0, was significantly enhanced compared to DeepSSETracer 1.1. On a set of 77 test cases, the weighted average per-voxel F1 score increased from 62.1% to 70.3% for helix detection, and from 47.8% to 62.5% for β-sheet detection. While each of the five modifications in the network enhanced the …

Copula-Based Bayesian Model For Detecting Differential Gene Expression, Prasansha Liyanaarachchi, N. Rao Chaganty

Mathematics & Statistics Faculty Publications

Deoxyribonucleic acid, more commonly known as DNA, is a fundamental genetic material in all living organisms, containing thousands of genes, but only a subset exhibit differential expression and play a crucial role in diseases. Microarray technology has revolutionized the study of gene expression, with two primary types available for expression analysis: spotted cDNA arrays and oligonucleotide arrays. This research focuses on the statistical analysis of data from spotted cDNA microarrays. Numerous models have been developed to identify differentially expressed genes based on the red and green fluorescence intensities measured using these arrays. We propose a novel approach using a Gaussian …

Hidden Markov Model For Identifying Local Variants In Human Genomes Using Simulated Data, Scott Mccallum

Electronic Theses and Dissertations

Identifying adaptive mutations in genetic data is challenging due to the low frequency of occurrence of such events, and because signatures of selection are intertwined with the footprints of various other evolutionary forces that shape our genomes. Even when a larger region appears to be under selection, genomic sites that are linked to adaptive mutations have similar statistical signals, and thus can obfuscate the identification of the actual adaptive mutation. The new method described here uses a Hidden Markov Model that allows for classification of neutral, linked, and sweep (adaptive mutation) genomic sites. This model is general and can be …

From Sampling To Simulating: Single-Cell Multiomics In Systems Pathophysiological Modeling, Alexandra Manchel, Michelle M. Gee, Rajanikanth Vadigepalli

Department of Pathology, Anatomy, and Cell Biology Faculty Papers

As single-cell omics data sampling and acquisition methods have accumulated at an unprecedented rate, various data analysis pipelines have been developed for the inference of cell types, cell states and their distribution, state transitions, state trajectories, and state interactions. This presents a new opportunity in which single-cell omics data can be utilized to generate high-resolution, high-fidelity computational models. In this review, we discuss how single-cell omics data can be used to build computational models to simulate biological systems at various scales. We propose that single-cell data can be integrated with physiological information to generate organ-specific models, which can then be …

Timigp: A Computational Framework To Determine The Tumor Immune Microenvironment Associated With Prognosis And Immunotherapy Response, Chenyang Li

Dissertations and Theses (Open Access)

Accumulating evidence has suggested that the tumor immune microenvironment (TIME) drastically impacts cancer patients’ clinical outcomes, including prognosis and immunotherapy response. However, understanding TIME remains challenging due to its complexity and heterogeneity. In this dissertation, we introduce TimiGP (Tumor Immune Microenvironment Illustration based on Gene Pairs), a computational framework designed to address this challenge. Leveraging single-cell RNA-seq (scRNA-seq) and bulk gene expression data alongside clinical information, TimiGP constructs a cell-cell interaction network that elucidates the relationship between immune cell function and relevant clinical outcomes, such as prognosis and treatment response. With immunological insights, these cell-cell interactions also facilitate the development …

Exercise Therapy Rescues Skeletal Muscle Dysfunction And Exercise Intolerance In Cardiometabolic Hfpef, Heather Quiriarte, Robert C. Noland, James E. Stampley, Gregory Davis, Zhen Li, Eunhan Cho, Youyoung Kim, Jake Doiron, Guillaume Spielmann, Sujoy Ghosh, Sanjiv J. Shah, Brian A. Irving, David J. Lefer, Timothy D. Allerton

School of Graduate Studies Faculty Publications

Exercise intolerance, a hallmark of heart failure with preserved ejection fraction (HFpEF) exacerbated by obesity, involves unclear mechanisms related to skeletal muscle metabolism. In a “2-hit” model of HFpEF, we investigated the ability of exercise therapy (voluntary wheel running) to reverse skeletal muscle dysfunction and exercise intolerance. Using state-of-the-art metabolic cages and a multiomic approach, we demonstrate exercise can rescue dysfunctional skeletal muscle lipid and branched-chain amino acid oxidation and restore exercise capacity in mice with cardiometabolic HFpEF. These results underscore the importance of skeletal muscle metabolism to improve exercise intolerance in HFpEF.

A Pan-Cancer Single-Cell Analysis Of Intratumoral Copy Number Diversity And Evolution, Hanghui Ye

Dissertations and Theses (Open Access)

Aneuploidy is a hallmark of human cancers, with many copy number aberrations (CNAs) being associated with disease progression. Previous studies have revealed extensive inter-patient heterogeneity (IPH) in copy number profiles. However, the extent of intratumoral heterogeneity (ITH) and its evolutionary dynamics remain poorly understood.

To address these gaps, we developed Acoustic Cell Tagmentation (ACT), an advanced single-cell single-molecule DNA sequencing (scDNA-seq) technology, to resolve the copy number substructure of human tumors and investigate the evolution of aneuploidy. Applying ACT to eight triple-negative breast cancer (TNBC) patients to profile 9,765 aneuploid tumor cells, we discovered that following initial punctuated copy number …

Investigating The Rare Aneuploid Cells In Normal Breast And Therapeutic Resistance In Triple Negative Breast Cancer Using Single Cell Genomics, Yiyun Lin

Dissertations and Theses (Open Access)

Aneuploid epithelial cells are common in breast cancer, however their presence in normal breast tissues is not well understood. To address this question, we applied single cell DNA sequencing to profile copy number alterations (CNAs) in 83,206 epithelial cells from breast tissues of 49 healthy women and single cell DNA&ATAC co-assays to 19 women. Our data shows that all women harbored rare aneuploid epithelial cells (median 3.19%) that increased with age. Many aneuploid epithelial cells (median 82.22%) in normal breast tissues underwent clonal expansions and harbored CNAs reminiscent of invasive breast cancers (gains of 1q, losses of 10q, 16q and …

Quantifying Effects Of Partial Genetic Backgrounds To Decode Genetic Drivers Of Clinical Phenotypes, Rini Pauly

All Dissertations

Understanding partial genetic backgrounds illuminates the genetic architecture of complex traits and diseases, revealing how diverse genetic backgrounds contribute to phenotypic diversity. With this approach we could advance personalized medicine by identifying population-specific variants affecting drug metabolism, tailoring medical treatments to individual genetic profiles. Additionally, it offers evolutionary insights into human history, shedding light on past migrations and the spread of genetic traits. This research leverages cutting-edge statistical genetics techniques and novel machine learning approaches to efficiently analyze extensive population genomic datasets, distilling complex admixture signals into meaningful genetic markers.

The study introduces Admix-AI, an innovative convolutional neural network-based tool …

Multi-Organ Gene Expression Analysis And Network Modeling Reveal Regulatory Control Cascades During The Development Of Hypertension In Female Spontaneously Hypertensive Rat, Eden Hornung, Sirisha Achanta, Alison Moss, James S. Schwaber, Rajanikanth Vadigepalli

Department of Pathology, Anatomy, and Cell Biology Faculty Papers

Hypertension is a multifactorial disease with stage-specific gene expression changes occurring in multiple organs over time. The temporal sequence and the extent of gene regulatory network changes occurring across organs during the development of hypertension remain unresolved. In this study, female spontaneously hypertensive (SHR) and normotensive Wistar Kyoto (WKY) rats were used to analyze expression patterns of 96 genes spanning inflammatory, metabolic, sympathetic, fibrotic, and renin-angiotensin (RAS) pathways in five organs, at five time points from the onset to established hypertension. We analyzed this multi-dimensional dataset containing ~15,000 data points and developed a data-driven dynamic network model that accounts for …