Biology subjects

Stuart, J.

Publications and source records attributed to Stuart, J..

14 recordsLinked to original sources

bioRxiv2024

Transformer-based modeling of Clonal Selection and Expression Dynamics (TraCSED) reveals resistance signatures in breast cancer

Understanding transcriptional heterogeneity in cancer cells and its implication for treatment response is critical to identify how resistance occurs and may be targeted. Such heterogeneity can be captured by in vitro studies through clonal barcoding methods. We present TraCSED (Transformer-based modeling of Clonal Selection and Expression Dynamics), a dynamic deep learning approach for modeling clonal selection. Using single-cell gene expression and the fitness of barcoded clones, TraCSED identifies interpretable gene programs and the timepoints at which they are associated with clonal selection. When applied to cells treated with either giredestrant, an estrogen receptor (ER) antagonist and degrader, or palbociclib, a CDK4/6 inhibitor, time-dependent resistance pathways are revealed. For example, ER activity is associated with positive selection around day four under palbociclib treatment and this adaptive response can be suppressed by combining the drugs. Yet, in the combination treatment, one clone still emerged. Clustering based on partial least squares regression found that high baseline expression of both SNHG25 and SNCG genes was the primary marker of positive selection to co-treatment and thus potentially associated with innate resistance - an aspect that traditional differential analysis methods missed. In conclusion, TraCSED enables associating pathways with phenotypes in a time-dependent manner from scRNA-seq data.

cancer biology↗

bioRxiv2023

A Spatiotemporal and Machine-Learning Platform Accelerates the Manufacturing of hPSC-derived Esophageal Mucosa

Human pluripotent stem cell-derived tissue engineering offers great promise in designer cell-based personalized therapeutics. To harness such potential, a broader approach requires a deeper understanding of tissue-level interactions. We previously developed a manufacturing system for the ectoderm-derived skin epithelium for cell replacement therapy. However, it remains challenging to manufacture the endoderm-derived esophageal epithelium, despite both possessing similar stratified structure. Here we employ single cell and spatial technologies to generate a spatiotemporal multi-omics cell atlas for human esophageal development. We illuminate the cellular diversity, dynamics and signal communications for the developing esophageal epithelium and stroma. Using the machine-learning based Manatee, we prioritize the combinations of candidate human developmental signals for in vitro derivation of esophageal basal cells. Functional validation of the Manatee predictions leads to a clinically-compatible system for manufacturing human esophageal mucosa. Our approach creates a versatile platform to accelerate human tissue manufacturing for future cell replacement therapies to treat human genetic defects and wounds.

developmental biology↗

bioRxiv2023

Associating Transcription Factors to Single-Cell Trajectories with DREAMIT

Trajectory methods have enabled the organization of cells into contiguous cellular changes from their transcriptional profiles measured by single cell RNA sequencing. Few methods enable investigating the implied gene regulatory network dynamics from the cell transitions between and along trajectory branches. In particular, there remains an opportunity to develop methods that leverage the predicted "pseudotime" orderings of cells to reveal transcription factor (TF) dynamics. Here we present DREAMIT (Dynamic Regulation of Expression Across Modules in Inferred Trajectories), a novel framework developed to detect patterns of TF activity along single-cell trajectory branches. It detects significant TF-target associations using a relational enrichment approach. Using a benchmark representing several different tissues, the method was found to have increased tissue-specific sensitivity and specificity over competing approaches. To illustrate the utility of the approach, we apply it to the analysis of a peripheral blood mononucleocyte dataset and discuss several examples of TF networks associated with monocytes and erythrocytes that reveal potential causal relationships among TFs. In summary, DREAMIT provides a useful tool for uncovering potential TF-to-target gene regulatory mechanisms associated with the cell-to-cell transitions predicted by trajectory inference methods.

bioinformatics↗

bioRxiv2021

Patient Informed Domain Adaptation Improves Clinical Drug Response Prediction

In-silico modeling of patient clinical drug response (CDR) promises to revolutionize personalized cancer treatment. State-of-the-art CDR predictions are usually based on cancer cell line drug perturbation profiles. However, prediction performance is limited due to the inherent differences between cancer cell lines and primary tumors. In addition, current computational models generally do not leverage both chemical information of a drug and a gene expression profile of a patient during training, which could boost prediction performance. Here we develop a Patient Adapted with Chemical Embedding (PACE) dual convergence deep learning framework that a) integrates gene expression along with drug chemical structures, and b) is adapted in an unsupervised fashion by primary tumor gene expression. We show that PACE achieves better discrimination between sensitive and resistant patients compared to the state-of-the-art linear regularized method (9/12 VS 3/12 drugs with available clinical outcomes) and alternative methods.

bioinformatics↗

bioRxiv2021

Mitochondrial and metabolic remodeling in human skin fibroblasts in response to glucose availability

Cell culture conditions highly influence cell metabolism in vitro. This is relevant for preclinical assays, for which fibroblasts are an interesting cell model, with applications in regenerative medicine, diagnostics and therapeutic development for personalized medicine as well as in the validation of ingredients for cosmetics. Given these cells short lifespan in culture, we aimed to identify the best cell culture conditions and promising markers to study mitochondrial health and stress in Normal Human Dermal Fibroblasts (NHDF). We tested the effect of reducing glucose concentration in the cell medium from high glucose (HGm) to a more physiological level (LGm), or its complete removal and replacement by galactose (OXPHOSm), always in the presence of glutamine and pyruvate. We have demonstrated that only with OXPHOSm it was possible to observe the selective inhibition of mitochondrial ATP production. This reliance on mitochondrial ATP was accompanied by changes in oxygen consumption rate (OCR) and extracellular acidification rate (ECAR), oxidation of citric acid cycle substrates, fatty acids, lactate and other substrates, mitochondrial network extension and polarization and changes in several key transcripts related to energy metabolism. We also evaluated the relevance of galactose, glutamine and pyruvate for OXPHOS stimulation, by comparing OCR and ECAR in the presence or absence of these substrates. Galactose and pyruvate seem to be important, but redundant, to promote OXPHOS, whereas glutamine was essential. We concluded that LGm does not promote significant metabolic changes but the short-term adaptation to OXPHOSm is ideal for studying mitochondrial health and stress in NHDF. Author ContributionsCC, SAP, SLCP and IMS performed experiments. TCO and PJO designed research and acquired funding. JS, and OB analyzed data. CC and TCO analyzed data and wrote the paper. All authors contributed to the final version of the manuscript.

cell biology↗

bioRxiv2021

Dual RNA-Seq analysis of SARS-CoV-2 correlates specific human transcriptional response pathways directly to viral expression

The SARS-CoV-2 pandemic has challenged humankinds ability to quickly determine the cascade of health effects caused by a novel infection. Even with the unprecedented speed at which vaccines were developed and introduced into society, identifying therapeutic interventions and drug targets for patients infected with the virus remains important as new strains of the virus may evolve, or future coronaviruses may emerge, that are resistant to current vaccines. The application of transcriptomic RNA sequencing of infected samples may shed new light on the pathways involved in viral mechanisms and host responses. We describe the application of "dual RNA-seq" analysis to consider both the host and pathogen transcriptomes simultaneously, to investigate for the first time the co-regulation of human and SARS-CoV-2 genes. Together with differential expression analysis, we describe the tissue specificity of SARS-CoV-2 expression, an inferred lipopolysaccharide response, and co-regulation of CXCLs, SPRRs, S100s with SARS-CoV-2 expression. Lipopolysaccharide response pathways in particular offer promise for future therapeutic research and the prospect of subgrouping patients based on chemokine expression that may help explain the vastly different reactions patients have to infection. Taken together these findings illuminate previously unappreciated SARS-CoV-2 expression signatures, identify new therapeutic considerations, and contribute a pipeline for studying multi-transcriptome systems.

bioinformatics↗

bioRxiv2020

Biological network-inspired interpretable variational autoencoder

Deep learning architectures such as variational autoencoders have revolutionized the analysis of transcriptomics data. However, the latent space of these variational autoencoders offers little to no interpretability. To provide further biological insights, we introduce a novel sparse Variational Autoencoder architecture, VEGA (Vae Enhanced by Gene Annotations), whose decoder wiring is inspired by a priori characterized biological abstractions, providing direct interpretability to the latent variables. We demonstrate the interpretability and flexibility of VEGA in diverse biological contexts, by integrating various sources of biological abstractions such as pathways, gene regulatory networks and cell type identities in the latent space of our model. We show that our model could recapitulate the mechanism of cellular-specific response to treatments, the status of master regulators as well as jointly investigate the cell type and cellular state identity in developing cells. We envision the approach could serve as an explanatory biological model in contexts such as development and drug treatment experiments.

bioinformatics↗

bioRxiv2020

Towards Inferring Nanopore Sequencing Ionic Currents from Nucleotide Chemical Structures

The characteristic ionic currents of nucleotide kmers are commonly used in analyzing nanopore sequencing readouts. We present a graph convolutional network-based deep learning framework for predicting kmer characteristic ionic currents from corresponding chemical structures. We show such a framework can generalize the chemical information of the 5-methyl group from thymine to cytosine by correctly predicting 5-methylcytosine-containing DNA 6mers, thus shedding light on the de novo detection of nucleotide modifications.

bioinformatics↗

bioRxiv2020

Highly Accurate Cancer Phenotype Prediction with AKLIMATE, a Stacked Kernel Learner Integrating Multimodal Genomic Data and Pathway Knowledge

Advancements in sequencing have led to the proliferation of multi-omic profiles of human cells under different conditions and perturbations. In addition, several databases have amassed information about pathways and gene "signatures" - patterns of gene expression associated with specific cellular and phenotypic contexts. An important current challenge in systems biology is to leverage such knowledge about gene coordination to maximize the predictive power and generalization of models applied to high-throughput datasets. However, few such integrative approaches exist that also provide interpretable results quantifying the importance of individual genes and pathways to model accuracy. We introduce AKLI-MATE, a first kernel-based stacked learner that seamlessly incorporates multi-omics feature data with prior information in the form of pathways for either regression or classification tasks. AKLIMATE uses a novel multiple-kernel learning framework where individual kernels capture the prediction propensities recorded in random forests, each built from a specific pathway gene set that integrates all omics data for its member genes. AKLIMATE outperforms state-of-the-art methods on diverse phenotype learning tasks, including predicting microsatellite instability in endometrial and colorectal cancer, survival in breast cancer, and cell line response to gene knockdowns. We show how AKLIMATE is able to connect feature data across data platforms through their common pathways to identify examples of several known and novel contributors of cancer and synthetic lethality.

bioinformatics↗

bioRxiv2019

A TBX5 dosage-sensitive gene regulatory network for human congenital heart disease

Haploinsufficiency of transcriptional regulators causes human congenital heart disease (CHD). However, underlying CHD gene regulatory network (GRN) imbalances are unknown. Here, we define transcriptional consequences of reduced dosage of the CHD-linked transcription factor, TBX5, in individual cells during cardiomyocyte differentiation from human induced pluripotent stem cells (iPSCs). We discovered highly sensitive dysregulation of TBX5-dependent pathways-- including lineage decisions and genes associated with cardiomyocyte function and CHD genetics--in discrete subpopulations of cardiomyocytes. GRN analysis identified vulnerable nodes enriched for CHD genes, indicating that cardiac network stability is sensitive to TBX5 dosage. A GRN-predicted genetic interaction between Tbx5 and Mef2c was validated in mouse, manifesting as ventricular septation defects. These results demonstrate exquisite sensitivity to TBX5 dosage by diverse transcriptional responses in heterogeneous subsets of iPSC-derived cardiomyocytes. This predicts candidate GRNs for human CHDs, with implications for quantitative transcriptional regulation in disease.

developmental biology↗

bioRxiv2019

Exploring Integrative Analysis using the BioMedical Evidence Graph

The analysis of cancer biology data involves extremely heterogeneous datasets including information from RNA sequencing, genome-wide copy number, DNA methylation data reporting on epigenomic regulation, somatic mutations from whole-exome or whole-genome analyses, pathology estimates from imaging sections or subtyping, drug response or other treatment outcomes, and various other clinical and phenotypic measurements. Bringing these different resources into a common framework, with a data model that allows for complex relationships as well as dense vectors of features, will unlock integrative analysis. We introduce a graph database and query engine for discovery and analysis of cancer biology, called the BioMedical Evidence Graph (BMEG). The BMEG is unique from other biological data graphs in that sample level molecular information is connected to reference knowledge bases. It combines gene expression and mutation data, with drug response experiments, pathway information databases and literature derived associations. The construction of the BMEG has resulted in a graph containing over 36M vertices and 29M edges. The BMEG system provides a graph query based API to enable analysis, with client code available for Python, Javascript and R, and a server online at bmeg.io. Using this system we have developed several forms of integrated analysis to demonstrate the utility of the system. The BMEG is an evolving resource dedicated to enabling integrative analysis. We have demonstrated queries on the system that illustrate mutation significance analysis, drug response machine learning, patient level knowledge base queries and pathway level analysis. We have compared the resulting graph to other available integrated graph systems, and demonstrated that it is unique in the scale of the graph and the type of data it makes available.\n\nHighlightsO_LIData resource connected extremely diverse set of cancer data sets\nC_LIO_LIGraph query engine that can be easily deployed and used on new datasets\nC_LIO_LIEasily installed python client\nC_LIO_LIServer online at bmeg.io\nC_LI\n\nSummaryThe analysis of cancer biology data involves extremely heterogeneous datasets including information. Bringing these different resources into a common framework, with a data model that allows for complex relationships as well as dense vectors of features, will unlock integrative analysis. We introduce a graph database and query engine for discovery and analysis of cancer biology, called the BioMedical Evidence Graph (BMEG). The construction of the BMEG has resulted in a graph containing over 36M vertices and 29M edges. The BMEG system provides a graph query based API to enable analysis, with client code available for Python, Javascript and R, and a server online at bmeg.io. Using this system we have developed several forms of integrated analysis to demonstrate the utility of the system.

cancer biology↗

bioRxiv2019

In vitro prototyping and rapid optimization of biosynthetic enzymes for cellular design

Microbial cell factories offer an attractive approach for production of biobased products. Unfortunately, designing, building, and optimizing biosynthetic pathways remains a complex challenge, especially for industrially-relevant, non-model organisms. To address this challenge, we describe a platform for in vitro Prototyping and Rapid Optimization of Biosynthetic Enzymes (iPROBE). In iPROBE, cell lysates are enriched with biosynthetic enzymes by cell-free protein synthesis and then metabolic pathways are assembled in a mix-and-match fashion to assess pathway performance. We demonstrate iPROBE with two examples. First, we tested and ranked 54 different pathways for 3-hydroxybutyrate production, improving in vivo production in Clostridium by 20-fold to 14.63 {+/-} 0.48 g/L and identifying a new biosynthetic route to (S)-(+)-1,3-butanediol. Second, we used iPROBE and data-driven design to optimize a 6-step n-butanol pathway, increasing titers 4-fold across 205 pathways, and showed strong correlation between cell-free and cellular performance. We expect iPROBE to accelerate design-build-test cycles for industrial biotechnology.

synthetic biology↗

bioRxiv2019

Cross-Species Alignment of Single Cell States with Biological Process Activity

The maintenance and transition of cellular states are controlled by biological processes. Here we present a gene set-based transformation of single cell RNA-Seq data into biological process activities that provides a robust description of cellular states. Moreover, as these activities represent species-independent descriptors, they facilitate the alignment of single cell states across different organisms.

systems biology↗

bioRxiv2017

Novel bacterial lineages associated with boreal moss species

Mosses are critical components of boreal ecosystems where they typically account for a large proportion of net primary productivity and harbor diverse bacterial communities that can be the major source of biologically-fixed nitrogen in these ecosystems. Despite their ecological importance, we have limited understanding of how microbial communities vary across boreal moss species and the extent to which local environmental conditions may influence the composition of these bacterial communities. We used marker gene sequencing to analyze bacterial communities associated with eight boreal moss species collected near Fairbanks, AK USA. We found that host identity was more important than site in determining bacterial community composition and that mosses harbor diverse lineages of potential N2- fixers as well as an abundance of novel taxa assigned to understudied bacterial phyla (including candidate phylum WPS-2). We performed shotgun metagenomic sequencing to assemble genomes from the WPS-2 candidate phylum and found that these moss-associated bacteria are likely anoxygenic phototrophs capable of carbon fixation via RuBisCo with an ability to utilize byproducts of photorespiration from hosts via a glyoxylate shunt. These results give new insights into the metabolic capabilities of understudied bacterial lineages that associate with mosses and the importance of plant hosts in shaping their microbiomes.

microbiology↗