Master of Science, The Ohio State University, 2010, Computer Science and Engineering

With the rapid development of high throughput gene expression profiling technology, molecule profiling has become a powerful tool to characterize disease subtypes and discover gene signatures. Most existing gene signature discovery methods apply statistical methods to select genes whose expression values can differentiate different subject groups. However, a drawback of these approaches is that the selected genes are not functionally related and hence cannot reveal biological mechanism behind the difference in the patient groups. Gene co-expression network analysis can be used to mine functionally related sets of genes that can be marked as potential biomarkers through survival analysis. We present an efficient heuristic algorithm EigenCut that exploits the properties of gene co-expression networks to mine functionally related and dense modules of genes. We apply this method to brain tumor (Glioblastoma Multiforme) study to obtain functionally related clusters. If functional groups of genes with predictive power on patient prognosis can be identified, insights on the mechanisms related to metastasis in GBM can be obtained and better therapeutical plan can be developed. We predicted potential biomarkers by dividing the patients into two groups based on their expression profiles over the genes in the clusters and comparing their survival outcome through survival analysis. We obtained 12 potential biomarkers with log-rank test p-values less than 0.01.

Committee: Kun Huang PhD (Committee Chair); Raghu Machiraju PhD (Committee Member) Subjects: Bioinformatics; Computer Science

Doctor of Philosophy, Case Western Reserve University, 2024, EECS - Computer and Information Sciences

With the rapid development of biotechnology that generate complex biological data, as well as graph machine learning algorithms to analyze these data, network-based analyses are becoming popular in modern biological research. Many network-based methods have been proposed to process biological data. Network embedding, which learns a representation of nodes into a low-dimensional space, has been a new learning paradigm in the studies of network analysis. Graph Representation learning algorithms are being commonly applied to a broad range of prediction tasks in systems biology. Mapping biological networks into a low-dimensional space enables efective application of machine learning methods in the downstream tasks. In this dissertation we present four new algorithms to compare and integrate biological networks, with a view to improving the reliability of graph machine learning algorithms on biological networks. These algorithms address inter-related problems to provide a comprehensive framework for network integration. Many real-world networks that are used in machine learning have multiple versions that come from diferent sources. For such networks, computation of Consensus Embeddings based on the node embeddings of individual versions can be useful for various reasons. GraphCan is a framework for computing canonical representations for biological networks using a similarity-based Graph Convolutional Network, and it integrates multiple node similarity measures to compute canonical node embeddings for a given network. Consensus embedding is used in our GraphCan model to integrate multiple node similarity measures to compute canonical node embeddings for a given network. GraphCan can be applied to diferent types of downstream tasks. BiGraphCan is an extension of GraphCan, and it aims to make bipartite predictions for understudied proteins using similarity networks and bipartite networks. We also explore network alignment problem by generalizing the Gromov-Wasserstein a (open full item for complete abstract)

Committee: Mehmet Koyuturk (Advisor); Rong Xu (Committee Member); Yinghui Wu (Committee Member); Jing Li (Committee Member) Subjects: Computer Science

Doctor of Philosophy, The Ohio State University, 2023, Computer Science and Engineering

In many biomedical contexts, multiple types of BDMs (e.g., metabolites, genes, proteins, chromatin states, and DNA methylation sites) associate with one another directly or indirectly in groups or chains to impact phenotype or outcome. Certain significant associations often help in data interpretation and novel hypotheses generation, motivating researchers to identify the most impactful groups of BDM associations between multiple types of data. However, many state-of-the-art models focus either on individual BDM associations independently of one another or implement black box predictors of outcome that are agnostic of BDM associations. Moreover, collection of multiple types of BDMs in a subject (i.e., multi-omics data) is not always feasible, motivating the need to infer one omic type of data from another. This dissertation tackles the related problems of (1) using inter-omics approaches to infer BDM types from other related BDM types in specific contexts, (2) finding groups of multi-omics data BDMs associated with outcome through multivariate statistical analysis and graph-based predictive models, and (3) interpreting groups of multi-omics data BDMs associated with outcome in a functional context using existing knowledge. This dissertation addresses the problem of using inter-omics approaches to infer BDM types from other related BDM types in two domains of note: (1) regulatory element annotation, and (2) protein abundance prediction. First, this dissertation introduces the Self Organizing Map with Variable Neighborhoods (SOM-VN), designed to annotate regulatory elements across whole human genomes using shapes found in chromatin accessibility assays. The novelty of SOM-VN is that, while most computational tools for annotating regulatory elements require a suite of resource-intensive experimental assays, SOM-VN uses only a single assay to annotate regulatory elements. SOM-VN is validated on chromatin accessibility assays from multiple H1, HeLa, A549, and GM12878 ce (open full item for complete abstract)

Committee: Raghu Machiraju (Advisor); Ewy Mathé (Advisor); Andrew Perrault (Committee Member); Rachel Kopec (Committee Member); Rachel Kelly (Committee Member) Subjects: Applied Mathematics; Artificial Intelligence; Bioinformatics; Biomedical Research; Biostatistics; Computer Science

Master of Science, The Ohio State University, 2023, Electrical and Computer Engineering

In this work, we performed a systematic literature review adopting a sensing → decision approach to identify the main primitives to establish an efficient end-end neuromorphic system. First, continuous-valued data (e.g., images or video frames) has to be encoded into spikes. Neuromorphic sensors can capture events and directly generate spikes instead of frames. However, for the purpose of this research, we implement encoding techniques to serve this purpose. We adopt the rate and latency (Time-To-First-Spike) coding techniques in this work. A Low-cost digital hardware implementation of both Rate and Latency coding is provided. Encoded spikes travel through synapses to reach processing neurons. Synaptic weights are learnable parameters for which a low-cost digital Spike-Time-Dependant-Plasticity (STDP) module is proposed and implemented in digital hardware. For the main processing element, we adopt the Leaky-Integrate-and-Fire (LIF) neuron model as a low-computational cost model, which is used in machine intelligence applications. In addition, a low-cost digital hardware implementation of the bio-plausible Izhikevich neuron model is provided for applications such as machine-brain interfaces. The work adopts a digital implementation of the advanced event representation (AER) communication scheme to establish the communication between different neurons within the system. Based on the components designed, an optimized 512-neuron 256K-synapse LIF-based digital neuromorphic processor is proposed and implemented, in which each neurosynaptic core can be configured as an encoding or processing neuron (LIF). We have also implemented a 1-1 software code for the processor, in which Pytorch and SNNtorch are used to build the spiking neural network, and surrogate gradients are used for training. An image classification task is used to evaluate the performance of the neuromorphic processor using the MNIST dataset, achieving an accuracy of 96.12% for 4-bit quantized synaptic we (open full item for complete abstract)

Committee: Waleed Khalil (Committee Member); Eslam Tawfik (Advisor) Subjects: Artificial Intelligence; Computer Engineering; Electrical Engineering; Engineering

Doctor of Philosophy, Case Western Reserve University, 2021, EECS - Computer and Information Sciences

Due to advances in experimental techniques, modern biological research provides an extensive and diverse set of data for computational analyses. At the genomic level, high throughput sequencing is capable of producing massive amounts of patient specific genetic information. Moving toward the fields of proteomics and cellular signaling, biological association data such as protein interactions are frequently studied though a graph theoretical model. These protein-protein interaction networks (PPIs) can then be extended by adding additional forms of network data such expression quantitative trait loci (eQTL), and disease associations, resulting in expansive heterogeneous networks. Furthermore, these networks are often tissue specific, all together representing a massive number of semantically useful network variations. This motivates the development of efficient compressed data structures and algorithms for working with versioned biological network data. In this dissertation I present algorithms and data-structures for efficiently compressing and querying biological data in real time. LinDen is a method for detecting epistatically interacting loci in genome wide association (GWAS) data. By hierarchically compressing loci according to their linkage disequilibrium between one another, it is possible to perform a highly accurate heuristic search for epistatically interacting locus pairs. VerTIoN is a compressed versioned sparse graph data-structure applied to the storage, retrieval, and integration of heterogeneous tissue specific networks including: protein interactions, eQTL interactions, and disease associations. I show that this method substantially improves the storage efficiency of tissue specific network data, while allowing fast decompression and composition. Finally, the work with VerTIoN is extended by utilizing it as the back-end of a multi-user versioned network query engine, enabling arbitrary on the fly version composition. To demonstrate the (open full item for complete abstract)

Committee: Mehmet Koyutürk (Committee Chair); Jing Li (Committee Member); Yinghui Wu (Committee Member); Rong Xu (Committee Member) Subjects: Bioinformatics; Computer Science

Doctor of Philosophy, The Ohio State University, 2009, Computer Science and Engineering

Many biological datasets can be effectively modeled as interaction networks where nodes represent biological entities of interest such as proteins, genes, or complexes and edges mimic associations among them. The study of these biological network structures can provide insight into many biological questions including the functional characterization of genes and gene products, the characterization of DNA-protein bindings, and the understanding of regulatory mechanisms. Therefore, the task of constructing biological interaction networks from raw data sets and exploiting information from these networks is critical, but is also fraught with challenges. First, the network structure is not always known in a priori; the structure should be inferred from raw and heterogeneous biological data sources. Second, biological networks are noisy (containing unreliable interactions) and incomplete (missing real interactions) which makes the task of extracting useful information difficult. Third, typically these networks have non-trivial topological properties (e.g., uneven degree distribution, small world) that limit the effectiveness of traditional knowledge discovery algorithms. Fourth, these networks are usually dynamic and investigation of their dynamics is essential to understand the underlying biological system. In this thesis, we address these issues by presenting a set of computational techniques that we developed to construct and analyze three specific types of biological interaction networks: protein-protein interaction networks, gene co-expression networks, and regulatory networks.

Committee: Parthasarathy Srinivasan (Advisor); Wang Yusu (Committee Member); Catalyurek Umit (Committee Member) Subjects: Bioinformatics; Computer Science