Search for: All records

Creators/Authors contains: "Murali, T M"

Note: When clicking on a Digital Object Identifier (DOI) number, you will be taken to an external site maintained by the publisher. Some full text articles may not yet be available without a charge during the embargo (administrative interval).
What is a DOI Number?

Some links on this page may take you to non-federal websites. Their policies may differ from this site.

  1. Advances in metagenomics have rapidly expanded viral discovery, revealing vast diversity across Earth’s virosphere. Yet most virus–host interactions—i.e., which viruses infect which hosts—remain unrecorded. Identifying these interactions is essential for anticipating zoonotic spillover events and advancing biomedical applications such as bacteriophage therapy. However, the sheer diversity of viruses and hosts makes comprehensive experimental mapping infeasible, motivating the need for computational approaches. Most existing prediction methods rely on supervised learning strategies that use sequence derived features, such as codon usage bias or k-mer frequencies, and do not model the coevolutionary processes that shape virus–host interactions. This limits their ability to generalize and the evolutionary interpretability of their predictions. We introduce CoEvoLink, a framework for predicting virus–host interactions that integrates sequence-based evidence with phylogenetic signal by explicitly modeling the coevolutionary histories of viruses and hosts. CoEvoLink infers likely but unobserved interactions by minimizing the number of evolutionary events required to explain them, yielding the most parsimonious interaction under a coevolutionary model. This formulation generalizes classical maximum parsimony, typically defined on a single phylogeny, by jointly optimizing parsimony across both virus and host phylogenies. Sequence-based information is incorporated by assigning a cost to each potential interaction that reflects its likelihood based on genomic features. By drawing a connection between computing parsimony on interaction matrices and maximum parsimony on phylogenetic networks, we derive a polynomial-time algorithm that balances parsimony with sequence-derived prediction cost. We demonstrate the effectiveness of CoEvoLink on simulated data under diverse coevolutionary models. Applying CoEvoLink, we identified putative bat hosts of betacoronaviruses that have not yet been cataloged in the VIRION database. On a benchmark derived from metagenomic sequencing data, we demonstrate that CoEvoLink improves the performance of existing phage-host prediction tools using cophylogenetic information. 
    more » « less
    Free, publicly-accessible full text available May 26, 2027
  2. The lack of high-quality ground truth datasets to train machine learning (ML) models impedes the potential of artificial intelligence for science (AI for science) to accelerate research. It is infeasible to manually review all relevant scientific publications and extract information from unstructured text to construct ground truth datasets. Scientific information extraction (SIE) from the literature using LLMs is emerging as a powerful approach to automate the creation of these datasets. However, existing LLM-based approaches and benchmarking studies for SIE focus on broad topics such as biomedicine and chemistry. These methods are limited to choice-based SIE tasks where the LLM is instructed to select the correct answer from several options provided in the prompt. Moreover, these tasks focus on extracting information from short and well-formatted text such as a sentence or an abstract. The potential of SIE methods in complex, open-ended tasks is considerably under-explored. In this study, we used a domain that has been virtually ignored in SIE, namely virology, to address these research gaps. We first designed a unique, open-ended SIE task of extracting mutations in a given virus that modify its interaction with the host. Next, we developed a new, multi-step retrieval augmented generation (RAG) framework called VILLA for SIE. In parallel, we curated a novel dataset of 629 mutations in ten influenza A virus proteins obtained from 463 scientific publications to serve as ground truth for our mutation extraction task. Finally, we demonstrated VILLA’s superior performance using a novel and comprehensive evaluation and comparison with vanilla RAG and other state-of-the art RAG- and agent-based tools involving both open and closed LLMs for SIE. 
    more » « less
    Free, publicly-accessible full text available August 9, 2027
  3. We propose a novel strategy for provenance tracing in random walk-based network diffusion algorithms, a problem that has been surprisingly overlooked in spite of the widespread use of diffusion algorithms in biological applications. Our path-based approach enables ranking paths by the magnitude of their contribution to each node’s score, offering insight into how information propagates through a network. Building on this capability, we introduce two quantitative measures: (i) path-based effective diffusion, which evaluates how well a diffusion algorithm leverages the full topology of a network, and (ii) diffusion betweenness, which quantifies a node’s importance in propagating scores. We applied our framework to SARS-CoV-2 protein interactors and human PPI networks. Provenance tracing of the Regularized Laplacian and Random Walk with Restart algorithms revealed that a substantial amount of a node’s score is contributed via multi-edge paths, demonstrating that diffusion algorithms exploit the non-local structure of the network. Analysis of diffusion betweenness identified proteins playing a critical role in score propagation; proteins with high diffusion betweenness are enriched with essential human genes and interactors of other viruses, supporting the biological interpretability of the metric. Finally, in a signaling network composed of causal interactions between human proteins, the top contributing paths showed strong overlap with COVID-19-related pathways. These results suggest that our path-based framework offers valuable insight into diffusion algorithms and can serve as a powerful tool for interpreting diffusion scores in a biologically meaningful context, complementing existing module- ornode-centric approaches in systems biology. The code is publicly available at https:// github.com/n-tasnina/provenance-tracing.git under the GNU General Public License v3.0. 
    more » « less
    Free, publicly-accessible full text available January 3, 2027
  4. ZooFlu is a database of influenza mutations contributing to mammalian adaptation and zoonosis potential. This is an iterative/updated database of influenza A virus mutations of zoonotic risk, manually curated, housed at GitHub (https://github.com/NSF-COMPASS-Center/zooflu). Mutations are organized for each major/minor IAV viral protein in Comma-Separated Values files (.csv). Details of the database are enumerated in the attached README file. 
    more » « less
  5. Abstract Synergistic drug combinations are often used to treat cancer. Experimental exploration of all possibilities is expensive. Deep learning (DL) offers a potential alternative for predicting drug pair synergy in specific cell lines. However, current methods often suffer from data leakage and lack systematic ablation studies. We propose SynVerse, a comprehensive evaluation framework featuring four data-splitting strategies to assess DL model generalizability and three ablation studies: module-based, feature shuffling, and a novel network-based approach to disentangle factors influencing performance. We evaluated sixteen models incorporating eight drug- and cell line-specific features, five preprocessing techniques, and two encoders. Our analysis revealed that no model outperformed a baseline using one-hot encoding. Biologically meaningful drug or cell line features and drug-drug interactions did not drive predictive performance. All models showed poor generalization to unseen drugs and cell lines. SynVerse highlights the need for substantial improvements before computational predictors can reliably support experimental and clinical settings. 
    more » « less
  6. Cheng, Jianlin (Ed.)
    Abstract MotivationA Multiple Sequence Alignment (MSA) contains fundamental evolutionary information that is useful in the prediction of structure and function of proteins and nucleic acids. The “Number of Effective Sequences” (NEFF) quantifies the diversity of sequences of an MSA. While several tools embed NEFF calculation with various options, none are standalone tools for this purpose, and they do not offer all the available options. ResultsWe developed NEFFy, the first software package to integrate all these options and calculate NEFF across diverse MSA formats for proteins, RNAs, and DNAs. It surpasses existing tools in functionality without compromising computational efficiency and scalability. NEFFy also offers per-residue NEFF calculation and supports NEFF computation for MSAs of multimeric proteins, with the capability to be extended to DNAs and RNAs. Availability and ImplementationNEFFy is released as open-source software under the GNU Public License v3.0. The source code in C ++ and a Python wrapper are available at https://github.com/Maryam-Haghani/NEFFy. To ensure users can fully leverage these capabilities, comprehensive documentation and examples are provided at https://Maryam-Haghani.github.io/NEFFy. Supplementary InformationSupplementary data are available at Bioinformatics online. 
    more » « less
  7. Large language models (LLMs) are increasingly touted as powerful tools for automating scientific information extraction. However, existing methods and tools often struggle with the realities of scientific literature: long-context documents, multi-modal content, and reconciling varied and inconsistent fine-grained information across multiple publications into standardized formats. These challenges are further compounded when the desired data schema or extraction ontology changes rapidly, making it difficult to re-architect or fine-tune existing systems. We present SciEx, a modular and composable framework that decouples key components including PDF parsing, multi-modal retrieval, extraction, and aggregation. This design streamlines on-demand data extraction while enabling extensibility and flexible integration of new models, prompting strategies, and reasoning mechanisms. We evaluate SciEx on datasets spanning three scientific topics for its ability to extract fine-grained information accurately and consistently. Our findings provide practical insights into both the strengths and limitations of current LLM-based pipelines. 
    more » « less
    Free, publicly-accessible full text available January 20, 2027
  8. Motivation: Molecular interaction networks are powerful tools for studying cellular functions. Integrating diverse types of networks enhances performance in downstream tasks such as gene module detection and protein function prediction. The challenge lies in extracting meaningful protein feature representations due to varying levels of sparsity and noise across these heterogeneous networks. Results: We propose ICoN, a novel unsupervised graph neural network model that takes multiple protein–protein association networks as inputs and generates a feature representation for each protein that integrates the topological information from all the networks. A key contribution of ICoN is exploiting a mechanism called “co-attention” that enables cross-network communication during training. The model also incorporates a denoising training technique, introducing perturbations to each input network and training the model to reconstruct the original network from its corrupted version. Our experimental results demonstrate that ICoN surpasses individual networks across three downstream tasks: gene module detection, gene coannotation prediction, and protein function prediction. Compared to existing unsupervised network integration models, ICoN exhibits superior performance across the majority of downstream tasks and shows enhanced robustness against noise. This work introduces a promising approach for effectively integrating diverse protein–protein association networks, aiming to achieve a biologically meaningful representation of proteins. Availability and implementation: The ICoN software is available under the GNU Public License v3 at https://github.com/Murali-group/ICoN. 
    more » « less
  9. {"Abstract":["This folder contains the datasets used for pre-training, fine-tuning, and evaluating HAVEN: Hierarchical Attention for Viral protEin-based host iNference. \n\nHAVEN is a viral protein language model pre-trained on 1.2 million protein sequences belonging to Viridae (viruses). It is fine-tuned to predict the hosts from which the viral protein sequence was sampled ("virus-host").\n\nThe viral protein sequences are downloaded from UniRef90. The host for the viral sequences are from the European Nucleotide Archive (ENA) maintained by European Molecular Biology Laboratory-European Bioinformatics Institute (EMBL-EBI)."]} 
    more » « less
  10. {"Abstract":["This folder contains the pre-trained and fine-tuned models of HAVEN: Hierarchical Attention for Viral protEin-based host iNference.HAVEN is a viral protein language model pre-trained on 1.2 million protein sequences belonging to Viridae (viruses). It is fine-tuned to predict the hosts from which the viral protein sequence was sampled ("virus-host"). HAVEN is based on the Bidirectional Encoder Representations from Transformers (BERT) architecture and is augmented with segmentation and hierarchical self-attention. This unique model design enables the model to accept sequences of highly diverse lengths.\n\n\n\nPlease refer the GitHub repository of HAVEN for information on using HAVEN.\n\n\n\nThe datasets used in the pre-training, fine-tuning, and evaluation of HAVEN are available at HAVEN Datasets."]} 
    more » « less