Biology subjects

Lathwal, S.

Publications and source records attributed to Lathwal, S..

5 recordsLinked to original sources

bioRxiv2025

DEEPGEOSEARCH: LLM-POWERED SCHEMALESS RETRIEVALFOR BIOMEDICAL DATA DISCOVERY

Public biomedical repositories contain extensive, valuable datasets, yet identifying datasets that precisely match specific study requirements remains inefficient. Conventional keyword- and schema-based systems frequently fall short when queries encompass multiple biological and experimental facets. To address this, we developed DeepGEOSearch, an LLM-powered, schema-less retrieval system that interprets dataset text directly rather than relying solely on predefined metadata fields. The system extracts key study attributes, harmonizes terminology across sources, and ranks datasets by contextual alignment with natural-language queries while providing verifiable evidence for each match. In this work, we applied DeepGEOSearch for the Gene Expression Omnibus (GEO). The framework is repository-agnostic and can integrate new metadata sources without manual relabeling. This design enables complex compositional queries that existing tools cannot support. In evaluations against strong baselines using a curated evaluation benchmark comprising a mix of query complexities, DeepGEOSearch achieved more than 90% precision and best recall, with the largest performance gains observed on complex, real-world queries. DeepGEOSearch consistently identifies relevant datasets overlooked by conventional search tools, accelerating dataset discovery and improving reuse of public biomedical data.

bioinformatics↗

bioRxiv2025

Global Hypomethylation in Cell-free DNA Enables Non-invasive Colorectal Cancer Screening: Results from a Retrospective Validation Study

Global DNA hypomethylation is a defining hallmark of colorectal cancer (CRC) but is poorly captured by existing cell-free DNA (cfDNA) technologies, which typically interrogate only a fraction of CpG sites and are biased toward CpG islands. We developed Asima Rev, an electrical-impedance cfDNA assay that can differentiate healthy individuals from those with cancer by measuring cfDNA aggregation patterns associated with methylation state, enabling functional detection of genome-wide hypomethylation. In a cohort of 46 treatment-naive CRC patients and 33 controls, Asima Rev achieved 96% sensitivity and 94% specificity, with longitudinal monitoring in six patients fully concordant with clinical outcomes. Interrogation of 11 public methylation array datasets showed that whole array analyses underestimate global changes. However, restricting analyses to OpenSea regions, where cfDNA is enriched, revealed patterns consistent with Asima Rev and a significant 5.8% global hypomethylation in CRC tissue compared to adjacent normal tissue. Hypomethylation was not observed in immune cell genomic DNA (gDNA), a major contributor to cfDNA, supporting a predominantly tumor-derived contribution to the observed cfDNA signal. Together, these results demonstrate that Asima Rev captures a cfDNA signal consistent with true global methylation loss and outperforms locus-specific assays by measuring structural consequences of pan-genomic epigenetic alterations.

bioinformatics↗

bioRxiv2020

Corna - An Open Source Python Tool For Natural Abundance Correction In Isotope Tracer Experiments

BackgroundStable isotope-based approaches are used in the field of metabolomics for quantification and identification of metabolites, discovery of new pathways and measurement of intracellular fluxes. In these experiments, often performed with mass spectrometry (MS), data must be corrected for natural abundance of isotopes. Various stand-alone tools with their own separate data formats and learning curves exist for correction of data collected at different resolutions, for tandem MS, and for different number of tracer elements. ResultsWe present a Python package, Corna, that combines natural abundance correction workflows for several experimental conditions and can be used as a one-stop-shop for stable isotope labeled experiments. We validate the algorithms in Corna with published tools, where available, and include new features, such as correction of two tracer elements, that are not yet implemented in any existing software application as per our knowledge. We also present the integration of Corna with an existing open source peak integration software. The integrated workflow can reduce processing times for a typical stable isotope based workflow from days to hours for a familiar user. ConclusionsAlgorithmic advancements have been keeping up with the developments in mass spectrometry technologies and have been the focus of most existing tools for natural abundance correction. However, in this high throughput era, it is also important to recognize user experience, and integrated and reproducible workflows. Corna has been written in Python and is designed for users who have access to large amounts of data from different kinds of experiments and want to integrate a natural abundance correction tool seamlessly in their pipelines. The latest version of Corna can be accessed at https://github.com/raaisakuk/NA_Correction.

bioinformatics↗

bioRxiv2019

Integrated Metabolomic and Transcriptomic Profiling Reveals Novel Activation-Induced Metabolic Networks in Human T cells

The targeting of metabolic pathways is emerging as an exciting new approach for modulating immune cell function and polarization states. In this study, carbon tracing and systems biology approaches integrating metabolomic and transcriptomic profiling data were used to identify adaptations in human T cell metabolism important for fueling pro-inflammatory T cell function. Results of this study demonstrate that T cell receptor (TCR) stimulation leads to a significant increase in glucose and amino acid metabolism that trigger downstream biosynthetic processes. Specifically, increased expression of several enzymes such as CTPS1, IL4I1, and ASL results in the reprogramming of amino acid metabolism. Additionally, the strength of TCR signaling resulted in different metabolic enzymes utilized by T cells to facilitate similar biochemical endpoints. Furthermore, this study shows that cyclosporine represses the pathways involved in amino acid and glucose metabolism, providing novel insights on the immunosuppressive mechanisms of this drug. To explore the implications of the findings of this study in clinical settings, conventional immunosuppressants were tested in combination with drugs that target metabolic pathways. Results showed that such combinations increased efficacy of conventional immunosuppressants. Overall, the results of this study provide a comprehensive resource for identifying metabolic targets for novel combinatorial regimens in the treatment of intractable immune diseases.

immunology↗

bioRxiv2018

Optimization of Experiment Design for Mass Spectrometric Isotopic Labeling Kinetics

Determination of metabolic fluxes by measurement of time-dependent sampling of isotopic enrichments during the administration of labeled substrates provides rich information. Because such experiments are resource-intensive and frequently push the limits of sensitivity of the measurement techniques, optimization of experiment design can improve feasibility with respect to financial and labor costs, time to completion, and increase precision and accuracy of the results. Here we used a previously published set of data acquired in cultured insulinoma cells to evaluate contributions to the sensitivity and variability of the rate of citrate synthase (CS). Specifically, we calculated changes in uncertainty in CS if sample times were dropped or new ones were added, and we observed that some sampling times can be dropped with little effect, while improvements can be made with a strategic choice of when to add samples. We measured the contributions of data sampled at different times on the sensitivity of CS, finding that CS had greater sensitivity at early time points. We tested the concept that if two estimated parameters are correlated significantly, then refining one might constrain the other. In this case, the rate of Beta-oxs was found to be correlated with CS, and narrower variability in Beta-ox did indeed improve the sensitivity of CS. The tests described here might be applied at the initial design stage and then after a pilot phase to improve sensitivities of targeted fluxes and the reduction of materials, time, labor, and other experimental resources. The correlation analyses can be used to consider what orthogonal measurements might be beneficial for further improvement of measurements. While this study used a specific example of a set of time-dependent kinetic isotopic measurements, the results illustrate some generalizable behaviors that can be tested in other experimental systems.

bioinformatics↗