One of the principal impediments to the broad success of conventional chemotherapy is poor delivery to and transport within the tumor microenvironment (TME), caused by irregular and leaky vasculature, the lack of functional lymphatics, and underscored by the overproduction of extracellular matrix (ECM) proteins such as collagen. Coupled with limited specificity, the high chemotherapeutic doses needed to effectively treat tumors often lead to unacceptable levels of damage to healthy tissues. Bacteria-based cancer therapy (BBCT) is an innovative alternative. Attenuated strains of species such as Salmonella Typhimurium have been shown to preferentially replicate in the TME, competing for cellular resources and imparting intrinsic and immune-mediated cytotoxic effects on cancer cells. Nevertheless, the immense successes observed in in vitro and immunocompromised murine models have not translated to the clinic, attributable to the lack of sufficient tumor colonization. Synthetic biology today enables the precision engineering of microbes with traits for improved survival, penetration, and replication in the TME, rationally optimizable through computational modeling. In this dissertation, we explore several facets of rationally engineering of bacteria toward augmenting bacterial penetration and retention in the TME. Namely, we (1) develop a novel assay to interrogate the neutrophil migratory response to pathogens and characterize the effects of modifying the molecular structure of the outer membrane (OM) of S. Typhimurium, (2) develop a mathematical model of bacterial intratumoral transport and growth and explore the effects of nutrient availability and the tumor ECM on colonization, (3) engineer bacteria that constitutively secrete collagenase and show significantly augmented transport in collagen hydrogels and collagen-rich tumor spheroids, and (4) develop computational models to explore control schemes for the programmed behavior of bacteria-based biohybrid systems, which will leverage the engineered bacteria to deliver therapeutics to the TME. Our work serves as the foundation for the logical and efficient design of the next generation of BBCTs. / Doctor of Philosophy / Cancer is one of the deadliest diseases facing our world today not because of a lack of effective medications in most cases, but because of our inability to target the diseased sites with those treatments. Many tumors lie in deep and sensitive regions that render them untouchable by direct physical means. Poor vascularization leads to only small fractions of toxic, systemically injected drugs being deposited in tumors. State-of-the-art treatments such as so-called "nano-medicines" that can target features of the diseased tissues and immunotherapies that train the immune system to attack tumor cells have made tremendous strides, but for many types of cancer, the underlying challenge of reaching cells far from blood vessels and targeting immunologically cold tumors remains. Bacteria-based cancer therapy (BBCT) presents an exciting opportunity to address these challenges. Based on microorganisms that can self-propel, proliferate, and display a preference for diseased tissues, their potential not only to carry chemotherapeutic payloads but also to elicit directly toxic or immunotherapeutic effects on cancer cells is clear from experimental work. Nevertheless, the same delivery and transport barriers facing other treatments, as well as immune-mediated clearance, have limited BBCTs' clinical success. Advances in synthetic biology and computational modeling today make the precision engineering of BBCT for traits that favor targeted cancer therapy a reality. The central hypothesis of this dissertation is that endowing tumor-targeting bacteria with a tissue-degrading enzyme has the potential to enhance tumor penetration and colonization. This dissertation work has led to development of computational and experimental frameworks for the design, testing, and optimization of BBCTs through direct quantitative assessment of the immune response, simulations to both optimize nutrient consumption for optimal growth and for programming genetic control strategies, and characterization of transport in tissue. Our work serves as a foundation for engineering "intelligent" BBCT.

Almost every precursor mRNA (pre-mRNA) in a eukaryotic organism undergoes splicing, in some cases resulting in the formation of more than one splice variant, a process called alternative splicing. RNA-Seq provides a major opportunity to capture the state of the transcriptome, which includes the detection of alternative spicing events. Alternative splicing is a highly regulated process occurring in a complex machinery called the spliceosome. In this dissertation, I focus on identification of different splice variants and splicing factors that are produced during Arabidopsis and soybean embryo development. I developed several data analysis pipelines for the detection and the functional characterization of active splice variants and splicing factors that arise during embryo development. The main goal of this dissertation was to identify transcriptional changes associated with specific stages of embryo development and infer possible associations between known regulatory genes and their targets. We identified several instances of exon skipping and intron retention as products of alternative splicing. The coding potential of the splice variants were evaluated using CodeWise. I developed CodeWise, a weighted support vector machine classifier to assess the coding potential of novel transcripts with respect to RNA secondary structure free energy, conserved domains, and sequence properties. We also examined the effect of alternative splicing on the domain composition of resulting protein isoforms. The majority of splice variants pairs encode proteins with identical domains or similar domains with truncation and in less than 10% of the cases alternative splicing results in gain or loss of a conserved domain. I constructed several possible regulatory networks that occur at specific stages of embryo development. In addition, in order to gain a better understanding of splicing regulation, we developed the concept of co-splicing networks, as a group of transcripts containing common RNA-binding motifs, which are co-expressed with a specific splicing factor. For this purpose, I developed a multi-stage analysis pipeline to integrate the co-expression networks with de novo RNA binding motif discovery at inferred splice sites, resulting in the identification of specific splicing factors and the corresponding cis-regulatory sequences that cause the production of splice variants. This approach resulted in the development of several novel hypotheses about the regulation of minor and major splicing in developing Arabidopsis embryos. In summary, this dissertation provides a comprehensive view of splicing regulation in Arabidopsis and soybean embryo development using computational analysis. / Ph. D.

The delivery of drugs into solid tumors is not trivial due to obstructions in the tumor microenvironment. Innovative drug delivery vehicles are currently being designed to overcome this challenge. In this research, computational fluid dynamics (CFD) simulations were used to evaluate the behavior of several drug delivery vectors in tumor capillaries—specifically motile bacteria, non-motile bacteria, and nanoparticles. Red blood cells, bacteria, and nanoparticles were imposed in the flow using the immersed boundary method. A human capillary model was developed using a novel method of handling deformable red blood cells (RBC). The capillary model was validated with experimental data from the literature. A stochastic model of bacteria motility was defined based on experimentally observed run and tumble behavior. The capillary and bacteria models were combined to simulate the intracapillary transport of bacteria. Non-motile bacteria and nanoparticles of 200 nm, 300 nm, and 405 nm were also simulated in capillary flow for comparison to motile bacteria. Motile bacteria tended to swim into the plasma layer near the capillary wall, while non-motile bacteria tended to get caught in the bolus flow between the RBCs. The nanoparticles were more impacted by Brownian motion and small scale fluid fluctuations, so they did not trend toward a single region of the flow. Motile bacteria were found to have the longest residence time in a 1 mm long capillary as well as the highest average radial velocity. This suggests motile bacteria may enter the interstitium at a higher rate than non-motile bacteria or nanoparticles of diameters between 200–405 nm. / Master of Science

<p dir="ltr">In the fields of computational biology and bioinformatics, the identification of novel disease biomarkers and therapeutic strategies, especially for cancer diseases, remains a crucial challenge. The advancement in computer science, particularly machine learning techniques, has greatly empowered the study of computational biology and bioinformatics for their unprecedented prediction power. This thesis explores how to utilize classic and advanced machine learning models to predict prognostic pathways, biomarkers, and therapeutics associated with cancers.</p><p dir="ltr">Firstly, this thesis presents a comprehensive overview of computational biology and bioinformatics, covering past milestones to current groundbreaking advancements, providing context for the research. A centerpiece of this thesis is the introduction of the Pathway Ensemble Tool and Benchmark, an original methodology designated for the unbiased discovery of cancer-related pathways and biomarkers. This toolset not only enhances the identification of crucial prognostic components distinguishing clinical outcomes in cancer patients but also guides the development of targeted drug treatments based on these signatures. Inspired by Benchmark, we extended the methodology to single-cell technologies and proposed scBenchmark and PathPCA, which provides insights into the potential and limitations of how novel techniques can benefit biomarker and therapeutic discovery. Next, the research progresses to the development of DREx, a deep learning model trained on large-scale transcriptome data, for predicting gene expression responses to drug treatments across multiple cell lines. DREx highlights the potential of advanced machine learning models in drug repurposing using a genomics-centric approach, which could significantly enhance the efficiency of initial drug selection.</p><p dir="ltr">The thesis concludes by summarizing these findings and highlighting their importance in advancing cancer-related biomarkers and drug discovery. Various computational predictions in this work have already been experimentally validated, showcasing the real-life impact of these methodologies. By integrating machine learning models with computational biology and bioinformatics, this research pioneers new standards for novel biomarker and therapeutics discovery.</p>

<p dir="ltr">Systems biology dynamic models, which are based on differential equations, offer a flexible and accurate framework to explain physiological properties emerging from complex biochem- ical or biological systems. These models enable explicit quantification and interpretation, allowing for simulation and perturbation analysis to study biological features and their inter- actions, as well as understanding system progression and convergence under various initial conditions. However, their application in human disease systems is limited due to unknown kinetics parameters under disease conditions and a reductionist paradigm that fails to cap- ture the complexity of diseases. Meanwhile, the advent of omics technologies provides high- resolution molecular measurements from single cells and spatially resolved samples, as well as comprehensive disease-specific molecular signatures from large patient cohorts. This wealth of data holds the promise for characterizing complex biological systems, necessitating ad- vanced systems biology models and computational tools that can harness multi-omics data to reliably depict biological processes. However, this endeavor faces the challenge of nonlinear relationships between omics data and the system’s dynamic properties, such as the global or local low-rank gene expression patterns across cell types and the nonlinear complexities within transcriptional regulatory networks revealed by single-cell RNA sequencing.</p><p dir="ltr">The overall goal of this report is to develop new computational frameworks, AI-empowered methods, and related mathematical theories to explicitly represent and approximate the dy- namics of complex biological systems by using biological omics data. Our aim is to unravel the intricacies of context-specific dynamic systems using multi-Omics data. Specifically, we solved two different but related computational tasks and enabled the first-of-its-kind methods to (1) identify local low-rank matrices from large omics data, and (2) a robust optimization strategy to approximate metabolic flux. Subsequently, we delve into the realm of data-driven and AI-powered systems biology, harnessing the power of statistical learning and artificial intelligence to approximate differential equations or their representations. This research en- deavor not only contributes to the advancement of subspace modeling but also offers insights into a wide array of complex phenomena across diverse domains, with profound implications for computational biology and beyond.</p>

The last decade has witnessed a tremendous increase in the amount of available biological data. Different technologies for measuring the genome, epigenome, transcriptome, proteome, metabolome, and microbiome in different organisms are producing large amounts of high-dimensional data every day. High-dimensional data provides unprecedented challenges and opportunities to gain a better understanding of biological systems. Unlike other data types, biological data imposes more constraints on researchers. Biologists are not only interested in accurate predictive models that capture complex input-output relationships, but they also seek a deep understanding of these models. In the last few years, deep models have achieved better performance in computational prediction tasks compared to other approaches. Deep models have been extensively used in processing natural data, such as images, text, and recently sound. However, application of deep models in biology is limited. Here, I propose to use deep models for output prediction, dimension reduction, and feature selection of biological data to get better interpretation and understanding of biological systems. I demonstrate the applicability of deep models in a domain that has a high and direct impact on health care. In this research, novel deep learning models have been introduced to solve pressing biological problems. The research shows that deep models can be used to automatically extract features from raw inputs without the need to manually craft features. Deep models are used to reduce the dimensionality of the input space, which resulted in faster training. Deep models are shown to have better performance and less variant output when compared to other shallow models even when an ensemble of shallow models is used. Deep models are shown to be able to process non-classical inputs such as sequences. Deep models are shown to be able to naturally process input sequences to automatically extract useful features. / Ph. D. / The world is generating more data than any time before. The abundance of data provides a great challenge and opportunity to get a better understanding of complex biological systems. The complexity of biological systems mandates better computational models that can make use the different types and formats of biological data. In the last few years, deep models have achieved better performance in computational prediction tasks compared to other approaches. Deep models have been extensively used in processing natural data, such as images, text, and recently sound. In this research, I show that deep learning can be applied to solve different biological problems that are directly related to human health. In this research, deep learning is used to predict which genes are essential for cancer cell survival. Deep learning is used to predict which drug combinations can work together to better treat cancer. Deep learning is used to predict whether two proteins are interacting with each other. This can be helpful for example in finding potential targets of viral proteins inside the human body.

Transcription factors (TFs) have been a focal point of molecular biology research for decades, with evolving methodologies offering progressively deeper insights into their critical roles in gene regulation. Recent advancements in experimental and computational techniques have significantly enhanced our understanding of TF functionality, yet this depth of knowledge varies widely across the spectrum of known TFs — from extensively characterized ones with quantitative binding affinity data to those scarcely studied or understood. In this work, we systematically carried out binding and expression experiments on all Escherichia coli TFs using a standardized computational pipeline to identify direct and indirect regulatory targets. We further leveraged our binding data to develop a novel biophysically motivated neural network capable of predicting TF-DNA binding affinity from DNA sequence. This approach allowed us to design binding sites with specified affinities, including those stronger than any sequence observed in nature, which we validate experimentally using an in vitro binding assay. We further optimized this assay to provide insight into complex TF binding regimes, where chemical signals can modulate TF binding affinity. Finally, we demonstrate the utility of systematically mapping TF binding sites through a case study on a previously thought dormant TF acquired from viral infection, revealing an unexpected phenotype where it can hijack the host cell. This work not only offers broad insights into the determinants of TF binding and regulation, but also provides a means to predictively engineer binding sites with desired affinity, while demonstrating the power of efficient data processing in uncovering intricate biological processes. / 2025-05-23T00:00:00Z

Yes / We present the Infobiotics Workbench (IBW), a user-friendly, scalable, and integrated computational environment for the computer-aided design of synthetic biological systems. It supports an iterative workflow that begins with specification of the desired synthetic system, followed by simulation and verification of the system in high- performance environments and ending with the eventual compilation of the system specification into suitable genetic constructs. IBW integrates modelling, simulation, verification and bicompilation features into a single software suite. This integration is achieved through a new domain-specific biological programming language, the Infobiotics Language (IBL), which tightly combines these different aspects of in silico synthetic biology into a full-stack integrated development environment. Unlike existing synthetic biology modelling or specification languages, IBL uniquely blends modelling, verification and biocompilation statements into a single file. This allows biologists to incorporate design constraints within the specification file rather than using decoupled and independent formalisms for different in silico analyses. This novel approach offers seamless interoperability across different tools as well as compatibility with SBOL and SBML frameworks and removes the burden of doing manual translations for standalone applications. We demonstrate the features, usability, and effectiveness of IBW and IBL using well-established synthetic biological circuits. / The work of S.K. is supported by EPSRC (EP/R043787/1). N.K., A.W., and B.B. acknowledge a Royal Academy of Engineering Chair in Emerging Technologies award and an EPSRC programme grant (EP/N031962/1).

Advancements in high-throughput technologies have led to an exponential increase in the generation of multi-modal data in computational biology. These datasets, comprising diverse biological measurements such as genomics, transcriptomics, proteomics, metabolomics, and imaging data, offer a comprehensive view of biological systems at various levels of complexity. However, integrating and analyzing such heterogeneous data present significant challenges due to differences in data modalities, scales, and noise levels. Another challenge for multi-modal analysis is the complex interaction network that the modalities share. Understanding the intricate interplay between different biological modalities is essential for unraveling the underlying mechanisms of complex biological processes, including disease pathogenesis, drug response, and cellular function. Machine learning algorithms have emerged as indispensable tools for studying multi-modal data in computational biology, enabling researchers to extract meaningful insights, identify biomarkers, and predict biological outcomes. In this dissertation, we first propose a multi-modal integration framework that takes two interconnected data modalities and their interaction network to iteratively update the modalities into new representations with better disease outcome predictive abilities. The deep learning-based model underscores the importance and performance gains achieved through the incorporation of network information into integration process. Additionally, a multi-modal framework is developed to estimate protein expression from mRNA and microRNA (miRNA) expressions, along with the mRNA-miRNA interaction network. The proposed network propagation model simulates in-vivo miRNA regulation on mRNA translation, offering a cost-effective alternative to experimental protein quantification. Analysis reveals that predicted protein expression exhibits a stronger correlation with ground truth protein expression compared to mRNA expression. Moreover, the effectiveness of integrative models is contingent upon the quality of input data modalities and the completeness of interaction networks, with missing values and network noise adversely affecting downstream tasks. To address these challenges, two multi-modal imputation models are proposed, facilitating the imputation of missing values in time series data. The first model allows the imputation of missing values in time series gene expression utilizing single nucleotide polymorphism (SNP) data for children at high risk of type 1 diabetes. The imputed gene expression allows us to predict the progression towards type 1 diabetes at birth with six years prediction horizon. Subsequently, a follow-up study introduces a generalized multi-modal imputation framework capable of imputing missing values in time series data using either another time series or cross-sectional data collected from the same set of samples. These models excel at imputation tasks, whether values are missing randomly or an entire time step in the series is absent. Additionally, leveraging the additional modality, they are able to estimate a completely missing time series without prior values. Finally, to mitigate noise in the interaction network, a link prediction framework for drug-target interaction prediction is developed. This study demonstrates exceptional performance in cold start predictions and investigates the efficacy of large language models for such predictions. Through a comprehensive review and evaluation of state-of-the-art algorithms, this dissertation aims to provide researchers with valuable insights, methodologies, and tools for harnessing the rich information embedded within multi-modal biological datasets.

An adequate ADME/PK (absorption, distribution, metabolism, excretion, pharmacokinetics) profile is an essential quality for a drug. As part of the drug discovery process, leads are iteratively designed and optimized in order to simultaneously satisfy various properties such as appropriate ADME/PK levels and high biological activity for a target. The drug discovery process can be accelerated by improving the likelihood that a designed compound fulfils the necessary pharmacologic properties, and thus reducing the number of needed iterations. A promising technique is de novo drug design, where molecules are computationally generated based on a set of desired attributes. Our project aimed to benchmark the effectiveness of the ANDROMEDA ADME/PK conformal prediction models in guiding the generation of compounds toward an area of chemical space with good ADME/PK properties. For this, we used the REINVENT reinforcement learning framework built by the Molecular AI team at AstraZeneca. Here, we integrated 4 out the 14 available ANDROMEDA models (fabs , fdiss, CLint and Vss) as oracles in the scoring component of the generative model. Oral bioavailability (F) is a secondary parameter that was computed with the help of the aforementioned models and fu(unbound fraction in plasma), and serves as the fifth ADME/PK oracle in our analysis. We aimed to rediscover DRD2 bioactives with a good ADME/PK profile. Our results show that the ANDROMEDA models have a slight influence on the predicted ADME/PK properties of the generated compounds. The results do not show an increased likelihood of generating DRD2 ligands in the case of the primary ANDROMEDA models. However, when using the oral bioavailability oracle, the sampling likelihood increases for some of the approved DRD2 ligands. In conclusion, the oral bioavailability ANDROMEDA model can be a promising option for guiding the generation of novel compounds towards an area of chemical space with good ADME/PK properties.