Search NASA⌕ Search
Home/Archive/DOE OSTI

DOE OSTI · 2572209

Harnessing Machine Learning for Agnostic Biodetection

Sandholtz, Sarah H. [Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)]·Valdes, Camilo [Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)]·Mulakken, Nisha [Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)]·Torres, Marisa W. [Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)]·Avila-Herrera, Aram [Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)]·Drocco, Jeffrey A. [Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)]·Martí, Jose Manuel [Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)]·Allen, Jonathan E. [Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)]·Tipnis, Uttara [Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)]·Jaing, Crystal J. [Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)]·Be, Nicholas A. [Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)]
View original source ↗Read source document ↗Export metadata ↓

Abstract

The United States’ current list-based approach to biodefense is limited because it considers only known biological agents. Alternatively, developing and adopting a system based on agent-agnostic signatures would enable detection and characterization of both known and novel agents, thereby engendering greater adaptability in the face of an evolving threat landscape. Machine learning (ML) could aid in such a transition, as it can recognize and encode highly complex patterns from multiple input data modalities and has already demonstrated success in many healthcare and defense applications. Functionalizing ML for environmental biodetection requires understanding current technical capabilities. In this article, we provide a systematic review of existing ML platforms and discuss anticipated development efforts needed to achieve effective ML-enabled, agnostic biodetection.

Explore related subjects

60 APPLIED LIFE SCIENCESbioelectricitybiotechnologygenetic engineeringregenerative medicine

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Sandholtz, Sarah H. [Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)], Valdes, Camilo [Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)], Mulakken, Nisha [Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)], Torres, Marisa W. [Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)], Avila-Herrera, Aram [Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)], Drocco, Jeffrey A. [Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)], Martí, Jose Manuel [Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)], Allen, Jonathan E. [Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)], Tipnis, Uttara [Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)], Jaing, Crystal J. [Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)], Be, Nicholas A. [Lawrence Livermore National Laboratory (LLNL), Livermore, CA (United States)]. 2025-05-30. Harnessing Machine Learning for Agnostic Biodetection. https://doi.org/10.1089/hs.2024.0075

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related reports

DOE OSTI2026

Metabolic flux and resource balance in the oleaginous yeast Rhodotorula toruloides

The yeast Rhodotorula toruloides is a promising bioproduction organism due to its high lipid yields and ability to grow on cheap and abundant substrates. Quantitative, systems-level assessment of its metabolic activity is accordingly merited. Resource-balance analysis (RBA) models capture not only reaction stoichiometry but also enzyme requirements for catalysis, providing valuable tools for understanding metabolic trade-offs and optimizing metabolic engineering strategies. Here, in this work, we present systems-level measurements of R. toruloides metabolic flux based on isotope tracing and metabolic flux analysis. In combination with new proteomic measurements, these flux data are used to parameterize a genome-scale resource balance model rtRBA. We find that S. cerevisiae and R. toruloides grow at nearly indistinguishable rates using similar biosynthetic but dramatically different central metabolic programs. R. toruloides consumes one-fifth as much glucose, which it metabolizes primarily via the pentose phosphate pathway and TCA cycle unlike primarily glycolysis in S. cerevisiae . Overall, across these two divergent yeasts, protein abundances aligned more closely than metabolic flux. Resource balance modeling of these metabolic programs predicts superior theoretical yields but lower productivities in R. toruloides than S. cerevisiae for industrial chemicals, highlighting the value of rapid glucose uptake for productivity but respiratory metabolism for yields.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2026

Opportunities for bioenergy crops to support transitions from irrigated agriculture and conserve the U.S. High Plains Aquifer

This study investigates potential economic and groundwater driven transitions from irrigated maize production—the dominant irrigated cropping system in the High Plains Aquifer (HPA) region—to alternative crops such as sorghum and switchgrass, two common bioenergy feedstocks. Unsustainable groundwater extraction in the U.S. High Plains presents critical challenges including reduced irrigation capacities, diminished crop yields, lower land values, and escalating energy costs. Using a spatially explicit optimization framework combined with comprehensive economic and hydrogeological data, we evaluate optimal land-use strategies and irrigation system investments over a 30-year planning period. Results indicate significant regional variability in the future economic viability of irrigated agriculture, driven by differences in aquifer recharge rates, groundwater availability, and market conditions. Nebraska and parts of northern Texas can sustain irrigated maize profitability due to relatively favorable groundwater conditions and lower land rents, respectively. In contrast, many portions of Kansas and southern Texas are more likely to transition to dryland agriculture within two decades. Colorado and New Mexico show potential for significant adoption of switchgrass production as an alternative biomass-based energy crop. Overall, over the 30-year horizon, our model implies that approximately 23% of currently irrigated maize area across the HPA region may transition to dryland farming under business-as-usual conditions. This transition is complemented by a threefold increase in non-irrigated sorghum production, from 0.20 to 0.77 Mt yr−1, indicating that groundwater-driven shifts in agricultural production may support a larger regional base for bioenergy feedstocks. The study also reveals opportunities for producers to optimize economic returns and biomass production potential.

60 APPLIED LIFE SCIENCES↗