Biology subjects

McDermott, J. E.

Publications and source records attributed to McDermott, J. E..

10 recordsLinked to original sources

bioRxiv2026

PathwaySeeker: Evidence-Grounded AI Reasoning over Organism-Specific Metabolic Networks

Metabolic activity is not an intrinsic property of an organism, but an emergent state shaped by environmental and experimental context. Despite recent advances in large language models (LLMs) and multi-omics profiling, current computational frameworks struggle to represent and reason over metabolism in a condition-specific manner. General-purpose AI systems operate on static, public biochemical knowledge, while multi-omics datasets capture dynamic measurements without a structured framework for mechanistic interpretation. As a result, metabolic networks remains analysis remains disconnected from the experimental states that define biological function. Here, we introduce PathwaySeeker, an evidence-grounded AI system for organism-specific metabolic network reasoning. PathwaySeeker reconstructs sample-specific metabolic graphs from integrated proteomic and metabolomic data, fine-tunes an LLM on the resulting graph structure, and verifies each reasoning step against the experimental graph through iterative hypothesis search, an approach we term Oracle-in-the-Loop inference. Every output claim carries explicit evidence provenance, distinguishing experimentally confirmed relationships from biochemically plausible hypotheses requiring validation. We demonstrate the system using multi-omics data from the non-model white-rot fungus Trametes versicolor, where PathwaySeeker recovers branched phenylpropanoid pathways and transparently stratifies confirmed reactions from testable extensions. Post-hoc thermodynamic analysis condition-specific metabolite dynamics support the biological feasibility of the reconstructed routes. By embedding experimental evidence provenance directly into language model-guided metabolic network reasoning, PathwaySeeker enables systematic differentiation between experimentally grounded knowledge and structured hypothesis, bridging frontier AI capabilities with organism-specific experimental evidence. Availability and ImplementationPathwaySeeker is available at https://github.com/pnnl/PathwaySeeker

bioinformatics↗

bioRxiv2026

VaLPAS: Leveraging variation in experimental multi-omics data to elucidate protein function

MotivationDespite continuing advances in sequencing and computational function determination, large parts of the studied gene, protein, and metabolite space remain functionally undetermined. Most function assignment is driven by homology searches and annotation transfer from known and extensively studied proteins but often fails to leverage available experimental omics data generated via technologies like mass-spectrometry. ResultsThe VaLPAS (Variation-Leveraged Phenomic Association Screen) framework is available as a Python package and provides a user-friendly platform for calculation of associations between expression patterns of genes or proteins in multi-omic datasets based on various statistical and learning methods. The goal of this approach is to shed light on the functional dark matter of protein space by elucidating previously unknown functions of molecules using guilt by association with molecules of known function. We present results demonstrating the utility of VaLPAS to identify high-confidence predictions for a subset of genes/proteins of unknown function in a previously published multi-omics dataset from the oleaginous yeast, Rhodotorula toruloides. AvailabilityVaLPAS is written in Python. The code is hosted on github (https://github.com/PNNL-Predictive-Phenomics/valpas/).

bioinformatics↗

bioRxiv2025

spammR: an R package designed for analysis and integration of spatial multi-omic measurements

SummarySpatial omics is a young and evolving field and as such shows rapid development of novel technologies and analysis methods to measure transcripts, proteins, metabolites, and post-translational modifications at high spatial resolution. These advances in technology have enabled the simultaneous generation of abundance profiles for multiple different omics types and associated microscopy imaging data, as well as their analysis in a spatial context. However, most analytical tools are designed for spatial transcriptomics platforms and are challenging to use in other contexts such as mass spectrometry-based measurements or metagenomics. To this end we present spammR (spatial analysis of multi-omics measurements in R), an R package that enables end-to-end analysis with a specific focus on mass-spectrometry derived spatial omics datasets with the goal of integration across multiple data types (e.g. sequencing, metabolites, proteins) within the same tissue sample. Availability and implementationspammR is implemented in R. The package is currently installable from GitHub (https://github.com/PNNL-CompBio/spammR).

bioinformatics↗

bioRxiv2025

Snekmer Learn/Apply: A kmer-based vector similarity approach to proteinclassification suitable for metagenomic datasets

Advances in whole genome sequencing have led to a rapid and ongoing increase in the amount of sequence data available, but 40-50% of known genes have no functional annotation and only 25-30% have specific functional annotations. Current functional annotation approaches typically rely on computationally expensive pairwise or multiple sequence alignments, preventing rapid development of models for novel protein functions and sometimes limiting methods to one ontology. Representation of sequence in short segments (kmers) has been used in many applications for nucleotide sequence, and more recently has been applied to protein sequence as well. We previously developed Snekmer, a tool which uses kmer patterns to develop alignment-free individual protein family models. Other approaches, such as MMSeqs2 and DIAMOND, use protein kmers as a fast filter to reduce search space for subsequent sequence alignment. Here, we describe a novel addition to the Snekmer tool which builds kmer libraries for protein families and uses those libraries to map functional annotations to new sequences. We first demonstrate that our method accurately applies TIGRFAMs annotations to protein fragments and to a low-sequence identity benchmark dataset, and further use it to annotate a set of drought stress associated soil and rhizosphere metagenome sequences with higher sensitivity towards several important protein function classes than that shown by HMMs. We have incorporated this workflow into Snekmer.

bioinformatics↗

bioRxiv2024

A functional screen for ubiquitin regulation identifies an E3 ligase secreted by Pseudomonas aeruginosa

Ubiquitin signaling controls many aspects of eukaryotic biology, including targeted protein degradation and immune defense. Remarkably, invading bacterial pathogens have adapted secreted effector proteins that hijack host ubiquitination to gain control over host responses. These ubiquitin-targeted effectors can exhibit, for example, E3 ligase or deubiquitinase activities, often without any sequence or structural homology to eukaryotic ubiquitin regulators. Such convergence in function poses a challenge to the discovery of additional bacterial virulence factors that target ubiquitin. To overcome this, we have developed a workflow to harvest natively secreted bacterial effectors and functionally screen them for ubiquitin regulatory activities. After benchmarking this approach on diverse ligase and deubiquitinase activities from Salmonella Typhimurium, Enteropathogenic Escherichia coli, and Shigella flexneri, we applied it to the identification of a cryptic E3 ligase activity secreted by Pseudomonas aeruginosa. We identified an unreported P. aeruginosa E3 ligase, which we have termed Pseudomonas Ub ligase 1 (PUL-1), that resembles none of the other E3 ligases previously established in or outside of the eukaryotic system. Importantly, in an animal model of P. aeruginosa infection, PUL-1 ligase activity plays an important role in regulating virulence. Thus, our workflow for the functional identification of ubiquitin-targeted effector proteins carries promise for expanding our appreciation of how host ubiquitin regulation contributes to bacterial pathogenesis.

microbiology↗

bioRxiv2023

Global Biogeography of the Soil Virosphere

Historically neglected by microbial ecologists, soil viruses are now thought to be critical to global biogeochemical cycles. However, our understanding of their global distribution, activities, and interactions with the soil microbiome remains limited. Here, we present the Global Soil Virus (GSV) Atlas, a comprehensive dataset compiled from 2,953 previously sequenced soil metagenomes and comprised of 616,935 uncultivated viral genomes (UViGs) and 38,508 unique viral operational taxonomic units (vOTUs). Rarefaction curves from the GSV Atlas indicate that most soil viral diversity remains unexplored, further underscored by high spatial turnover and low rates of shared vOTUs across samples. By examining genes associated with biogeochemical functions, we also demonstrate the viral potential to impact soil carbon and nutrient cycling. This study represents an extensive characterization of soil viral diversity and provides a foundation for developing testable hypotheses regarding the role of the virosphere in the soil microbiome and global biogeochemistry.

ecology↗

bioRxiv2023

Sorgoleone degradation by sorghum-associated bacteria; an opportunity for enforcing plant growth promotion

Metabolite exchange between plant roots and their associated rhizosphere microbiomes underpins plant growth promotion by microbes. Sorghum bicolor is a cereal crop that feeds animals and humans and is used for bioethanol production. Its root tips exude large amounts of a lipophilic benzoquinone called sorgoleone. Sorgoleone is an allelochemical that suppresses the growth of competing plant seedlings and is mineralized by microbes in soil. As an avenue to understand how sorghum and its root microbiome may be connected through root exudates, we identified the molecular determinants of microbial sorgoleone degradation and the distribution of this trait among microbes. We isolated and studied from sorghum-associated soils, three bacterial strains classified as Acinetobacter, Burkholderia, and Pseudomonas species that grow with sorgoleone as a sole carbon and energy source. The genomes of these strains were sequenced and subjected to transcriptomic and gene fitness analyses to identify candidate sorgoleone degradation genes. Follow up mutational analysis showed that sorgoleone catabolism is dependent on four contiguous genes that are conserved among the species we sequenced. Phylogenetic analysis of the sorgoleone degradation gene cluster showed that sorgoleone catabolism is enriched in sorghum-associated Streptomyces strains. The discovery of bacteria that grow on a compound like sorgoleone that is plant specific and not widely distributed in the environment, provides an opportunity to study how a plant exudate can enforce the development of a rhizosphere specific microbiome for the mutual benefit of plant and microbe. SignificanceThe grain crop sorghum exudes an herbicidal compound called sorgoleone from its root tips, which inhibits the growth of other plants. We isolated bacteria that grow on sorogleone and identified a cluster of bacterial genes required for sorogleone degradation that can be used as a biomarker for this trait. An approach to improve the production of crops in stressful conditions such as drought, is to encourage their association with plant growth promoting bacteria. Our discovery of sorgoleone degradation genes opens the door to engineering bacteria that receive benefit from sorghum in the form of a plant-specific growth substrate, and in return promote the growth of this crop.

microbiology↗

bioRxiv2022

MetaPredict: A machine learning-based tool for predicting metabolic modules in incomplete bacterial genomes

The reconstruction of complete microbial metabolic pathways using omics data from environmental samples remains challenging. Computational pipelines for pathway reconstruction that utilize machine learning methods to predict the presence or absence of KEGG modules in incomplete genomes are lacking. Here, we present MetaPathPredict, a software tool that incorporates machine learning models to predict the presence of complete KEGG modules within bacterial genomic datasets. Using gene annotation data and information from KEGG module databases, MetaPathPredict employs neural network and XGBoost stacked ensemble models to reconstruct and predict the presence of KEGG modules in a genome. MetaPathPredict can be used as a command line tool or as an R package, and both options are designed to be run locally or on a compute cluster. In our benchmarks, MetaPathPredict makes robust predictions of KEGG module presence within highly incomplete genomes.

bioinformatics↗

bioRxiv2021

Utilizing proteomics and phosphoproteomics to predict ex vivo drug sensitivity across genetically diverse AML patients

Acute Myeloid Leukemia (AML) affects 20,000 patients in the US annually with a five-year survival rate of approximately 25%. One reason for the low survival rate is the high prevalence of clonal evolution that gives rise to heterogeneous sub-populations of leukemia. This genetic heterogeneity is difficult to treat using conventional therapies that are generally based on the detection of a single driving mutation. Thus, the use of molecular signatures, consisting of multiple functionally related transcripts or proteins, in making treatment decisions may overcome this hurdle and provide a more effective way to inform drug treatment protocols. Toward this end, the Beat AML research program prospectively collected genomic and transcriptomic data from over 1000 AML patients and carried out ex vivo drug sensitivity assays to identify signatures that could predict patient-specific drug responses. The Clinical Proteomic Tumor Analysis Consortium is in the process of extending this cohort to collect proteomic and phosphoproteomic measurements from a subset of these patient samples to evaluate the hypothesis that proteomic signatures can robustly predict drug response in AML patients. We sought to examine this hypothesis on a sub-cohort of 38 patient samples from Beat AML with proteomic and drug response data and evaluate our ability to identify proteomic signatures that predict drug response with high accuracy. For this initial analysis we built predictive models of patient drug responses across 26 drugs of interest using the proteomics and phosphproteomics data. We found that proteomics-derived signatures provide an accurate and robust signature of drug response in the AML ex vivo samples, as well as related cell lines, with better performance than those signatures derived from mutations or mRNA expression. Furthermore, we found that in specific drug-resistant cell lines, the proteins in our prognostic signatures represented dysregulated signaling pathways compared to parental cell lines, confirming the role of the proteins in the signatures in drug resistance. In conclusion, this pilot study demonstrates strong promise for proteomics-based patient stratification to predict drug sensitivity in AML.

bioinformatics↗

bioRxiv2019

Untapped viral diversity in global soil metagenomes

Viruses outnumber every other biological entity on Earth, and soil viruses are particularly diverse compared to other habitats. However, we have limited understanding of soil viruses because of the tremendous variation in soil ecosystems and because of the lack of appropriate screening tools. Here, we determined the global distribution of more than 24,000 soil viral sequences and their potential hosts, including >1,600 sequences associated with giant viruses. The viral sequences, derived from 668 terrestrial metagenomes, greatly extend existing knowledge of soil viral diversity and viral biogeographical distribution. We screened these sequences to identify a suite of cosmopolitan auxiliary metabolic genes (AMGs) encoding enzymes involved in soil organic carbon decomposition across soil biomes. Additionally, we provide evidence for viral facilitation of multi-domain linkages in soils by locating a fungal chitosanase in bacteriophages, generating a new paradigm of how viruses can serve as exchange vectors of carbon metabolism across domains of life.

microbiology↗