Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “bioinformatics tool”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 127 records · Page 7
All sourcesNASA NTRSDOE OSTI
DOE OSTI2022

Metagenome-assembled genome extraction and analysis from microbiomes using KBase

Uncultivated Bacteria and Archaea account for the vast majority of species on Earth, but obtaining their genomes directly from the environment, using shotgun sequencing, has only become possible recently. In order to realize the hope of capturing Earth’s microbial genetic complement and to facilitate the investigation of the functional roles of specific lineages in a given ecosystem, technologies that accelerate the recovery of high-quality genomes are necessary. We present a series of analysis steps and data products for the extraction of high-quality metagenome-assembled genomes (MAGs) from microbiomes using the U.S. Department of Energy Systems Biology Knowledgebase (KBase) platform (http://www.kbase.us/). Overall, these steps take about a day to obtain extracted genomes when starting from smaller environmental shotgun read libraries, or up to about a week from larger libraries. In KBase, the process is end-to-end, allowing a user to go from the initial sequencing reads all the way through to MAGs, which can then be analyzed with other KBase capabilities such as phylogenetic placement, functional assignment, metabolic modeling, pangenome functional profiling, RNA-Seq and others. While portions of such capabilities are available individually from other resources, the combination of the intuitive usability, data interoperability and integration of tools in a freely available computational resource makes KBase a powerful platform for obtaining MAGs from microbiomes. While this workflow offers tools for each of the key steps in the genome extraction process, it also provides a scaffold that can be easily extended with additional MAG recovery and analysis tools, via the KBase software development kit (SDK).

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2024

Spatial Proteomics towards cellular Resolution

Introduction: Spatial biology is an emerging interdisciplinary field facilitating biological discoveries through the use of spatial omics technologies. Recent advancements in spatial transcriptomics, spatial genomics (e.g. genetic mutations and epigenetic marks), multiplexed immunofluorescence, and spatial metabolomics/lipidomics have enabled high-resolution spatial profiling of gene expression, genetic variation, protein expression, and metabolites/lipids profiles in tissue. These developments contribute to a deeper understanding of the spatial organization within tissue microenvironments at the molecular level. Areas covered: This report provides an overview of the untargeted, bottom-up mass spectrometry (MS)-based spatial proteomics workflow. It highlights recent progress in tissue dissection, sample processing, bioinformatics, and liquid chromatography (LC)-MS technologies that are advancing spatial proteomics toward cellular resolution. Expert opinion: The field of untargeted MS-based spatial proteomics is rapidly evolving and holds great promise. To fully realize the potential of spatial proteomics, it is critical to advance data analysis and develop automated and intelligent tissue dissection at the cellular or subcellular level, along with high-throughput LC-MS analyses of thousands of samples. In conclusion, achieving these goals will necessitate significant advancements in tissue dissection technologies, LC-MS instrumentation, and computational tools.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2024

An FPGA-based hardware accelerator supporting sensitive sequence homology filtering with profile hidden Markov models

Abstract Background Sequence alignment lies at the heart of genome sequence annotation. While the BLAST suite of alignment tools has long held an important role in alignment-based sequence database search, greater sensitivity is achieved through the use of profile hidden Markov models (pHMMs). Here, we describe an FPGA hardware accelerator, called HAVAC, that targets a key bottleneck step (SSV) in the analysis pipeline of the popular pHMM alignment tool, HMMER. Results The HAVAC kernel calculates the SSV matrix at 1739 GCUPS on a $$\sim$$ ∼ $3000 Xilinx Alveo U50 FPGA accelerator card, $$\sim$$ ∼ 227× faster than the optimized SSV implementation in nhmmer . Accounting for PCI-e data transfer data processing, HAVAC is 65× faster than nhmmer’s SSV with one thread and 35× faster than nhmmer with four threads, and uses $$\sim$$ ∼ 31% the energy of a traditional high end Intel CPU. Conclusions HAVAC demonstrates the potential offered by FPGA hardware accelerators to produce dramatic speed gains in sequence annotation and related bioinformatics applications. Because these computations are performed on a co-processor, the host CPU remains free to simultaneously compute other aspects of the analysis pipeline.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2020

An open-source high-content analysis workflow for CFTR function measurements using the forskolin-induced swelling assay

Abstract Motivation The forskolin-induced swelling (FIS) assay has become the preferential assay to predict the efficacy of approved and investigational CFTR-modulating drugs for individuals with cystic fibrosis (CF). Currently, no standardized quantification method of FIS data exists thereby hampering inter-laboratory reproducibility. Results We developed a complete open-source workflow for standardized high-content analysis of CFTR function measurements in intestinal organoids using raw microscopy images as input. The workflow includes tools for (i) file and metadata handling; (ii) image quantification and (iii) statistical analysis. Our workflow reproduced results generated by published proprietary analysis protocols and enables standardized CFTR function measurements in CF organoids. Availability and implementation All workflow components are open-source and freely available: the htmrenamer R package for file handling https://github.com/hmbotelho/htmrenamer; CellProfiler and ImageJ analysis scripts/pipelines https://github.com/hmbotelho/FIS_image_analysis; the Organoid Analyst application for statistical analysis https://github.com/hmbotelho/organoid_analyst; detailed usage instructions and a demonstration dataset https://github.com/hmbotelho/FIS_analysis. Distributed under GPL v3.0. Supplementary information Supplementary data are available at Bioinformatics online.

Hagemeijer, Marne C.↗
DOE OSTI2021

DNCON2_Inter: predicting interchain contacts for homodimeric and homomultimeric protein complexes using multiple sequence alignments of monomers and deep learning

Deep learning methods that achieved great success in predicting intrachain residue-residue contacts have been applied to predict interchain contacts between proteins. However, these methods require multiple sequence alignments (MSAs) of a pair of interacting proteins (dimers) as input, which are often difficult to obtain because there are not many known protein complexes available to generate MSAs of sufficient depth for a pair of proteins. In recognizing that multiple sequence alignments of a monomer that forms homomultimers contain the co-evolutionary signals of both intrachain and interchain residue pairs in contact, we applied DNCON2 (a deep learning-based protein intrachain residue-residue contact predictor) to predict both intrachain and interchain contacts for homomultimers using multiple sequence alignment (MSA) and other co-evolutionary features of a single monomer followed by discrimination of interchain and intrachain contacts according to the tertiary structure of the monomer. We name this tool DNCON2_Inter. Allowing true-positive predictions within two residue shifts, the best average precision was obtained for the Top-L/10 predictions of 22.9% for homodimers and 17.0% for higher-order homomultimers. In some instances, especially where interchain contact densities are high, DNCON2_Inter predicted interchain contacts with 100% precision. We also developed Con_Complex, a complex structure reconstruction tool that uses predicted contacts to produce the structure of the complex. Using Con_Complex, we show that the predicted contacts can be used to accurately construct the structure of some complexes. Our experiment demonstrates that monomeric multiple sequence alignments can be used with deep learning to predict interchain contacts of homomeric proteins.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2026

Exploring life’s hidden majority: microbial dark matter symposium highlights

The Microbial Dark Matter Symposium held on August 28–29, 2025, in Laguna Beach, Orange County, CA, convened a multidisciplinary group of scientists to address the vast unknowns in microbial life—from uncultured taxa and uncharacterized proteins to elusive viruses and spacefaring microbes. Set against a scenic coastal backdrop, the symposium highlighted advances in single-cell genomics, proximity ligation sequencing, and artificial intelligence-ready bioinformatics, while also probing the limits of microbial persistence, metabolism, and ecological distribution. Sessions explored microbial dark matter from multiple dimensions: cultivability, where new strategies are enabling recovery of elusive microbes; functional ambiguity, where metagenomic dark zones are illuminated by computational annotation; and genomic representation, where single-cell methods bridge gaps left by shotgun community sequencing. Researchers shared breakthroughs in identifying atmospheric microbiomes, “dark oxygen” production in groundwater ecosystems, and microbial survival on the International Space Station. The symposium emphasized integration of methods, disciplines, and ecosystems, advancing a collective push to illuminate the microbial dark matter on Earth and beyond. By highlighting emerging tools, pressing questions, and cross-domain insights, the symposium underscored the need for collaborative, open, and adaptive approaches to study the microbial unknown. The meeting marks a pivotal moment in microbiology, where cultivating knowledge of the uncultivated promises transformative understanding of life, everywhere.

Podar, Mircea [ORNL] (ORCID:0000000327760205)↗
DOE OSTI2020

RCSB Protein Data Bank tools for 3D structure-guided cancer research: human papillomavirus (HPV) case study

Abstract Atomic-level three-dimensional (3D) structure data for biological macromolecules often prove critical to dissecting and understanding the precise mechanisms of action of cancer-related proteins and their diverse roles in oncogenic transformation, proliferation, and metastasis. They are also used extensively to identify potentially druggable targets and facilitate discovery and development of both small-molecule and biologic drugs that are today benefiting individuals diagnosed with cancer around the world. 3D structures of biomolecules (including proteins, DNA, RNA, and their complexes with one another, drugs, and other small molecules) are freely distributed by the open-access Protein Data Bank (PDB). This global data repository is used by millions of scientists and educators working in the areas of drug discovery, vaccine design, and biomedical and biotechnology research. The US Research Collaboratory for Structural Bioinformatics Protein Data Bank (RCSB PDB) provides an integrated portal to the PDB archive that streamlines access for millions of worldwide PDB data consumers worldwide. Herein, we review online resources made available free of charge by the RCSB PDB to basic and applied researchers, healthcare providers, educators and their students, patients and their families, and the curious public. We exemplify the value of understanding cancer-related proteins in 3D with a case study focused on human papillomavirus.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2024

Signatures of Mollicutes-related endobacteria in publicly available Mucoromycota genomes

ABSTRACT Mucoromycota fungi and their Mollicutes-related endobacteria (MRE) are an ideal system for studying bacterial–fungal interactions and evolution due to the long-term and intimate nature of their interactions. However, methods for detecting MRE face specific challenges due to the poor representation of MRE in sequencing databases coupled with the high sequence divergence of their genomes, making traditional similarity searches unreliable. This has precluded estimations on the diversity of MRE associated with Mucoromycota. To determine the prevalence of previously undetected MRE in fungal genome sequences, we scanned 389 Mucoromycota genome assemblies available from the National Center for Biotechnology Information for the presence of MRE sequences using publicly available tools to map contigs from fungal assemblies to publicly available MRE genomes. We demonstrate a higher diversity of MRE genomes than previously described in Mucoromycota and a lack of cophylogeny between MRE and the majority of their fungal hosts. This supports the late invasion hypothesis regarding MRE acquisition across most of the examined fungal families. In contrast with other Mucoromycota lineages, MRE from the Gigasporaceae displayed some degree of cophylogeny with their hosts, which may indicate that horizontal transmission is restricted between members of this family or that transmission is strictly vertical. These results underscore the need for a refined process to capture sequencing data from potential fungal endosymbionts to discern their evolution and transmission. Screens of fungal genomes for MRE can help improve the quality of fungal genome assemblies while identifying new MRE lineages to further test hypotheses on their origin and evolution. IMPORTANCE Mollicutes-related endobacteria (MRE) are obligate intracellular bacteria found within Mucoromycota fungi. Despite their frequent detection, MRE roles in host functioning are still unknown. Comparative genomic investigations can improve our understanding of the impact of MRE on their fungal hosts by identifying similarities and differences in MRE genome evolution. However, MRE genomes have only been assembled from a small fraction of Mucoromycota hosts. Here, we demonstrate that MRE can be present yet undetected in publicly available Mucoromycota genome assemblies. We use these newfound sequences to assess the broader diversity of MRE and their phylogenetic relationships with respect to their hosts. We demonstrate that publicly available tools can be used to extract novel MRE sequences from assembled fungal genomes leading to insights on MRE evolution. This work contributes to a greater understanding of the fungal microbiome, which is crucial to improving knowledge on the dynamics and impacts of fungi in microbial ecosystems.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2022

Diversification of aminoacyl-tRNA synthetase activities via genomic duplication

Intricate evolutionary events enabled the emergence of the full set of aminoacyl-tRNA synthetase (aaRS) families that define the genetic code. The diversification of aaRSs has continued in organisms from all domains of life, yielding aaRSs with unique characteristics as well as aaRS-like proteins with innovative functions outside translation. Recent bioinformatic analyses have revealed the extensive occurrence and phylogenetic diversity of aaRS gene duplication involving every synthetase family. However, only a fraction of these duplicated genes has been characterized, leaving many with biological functions yet to be discovered. Here we discuss how genomic duplication is associated with the occurrence of novel aaRSs and aaRS-like proteins that provide adaptive advantages to their hosts. We illustrate the variety of activities that have evolved from the primordial aaRS catalytic sites. This precedent underscores the need to investigate currently unexplored aaRS genomic duplications as they may hold a key to the discovery of exciting biological processes, new drug targets, important bioactive molecules, and tools for synthetic biology applications.

noncanonical functions↗
DOE OSTI2026

Soil metagenomics umbrella narrative

Implementing accessible, authentic research experiences in introductory courses is challenging, particularly at institutions serving diverse student populations. To address this gap, we developed and deployed a Course-based Undergraduate Research Experience (CURE) focused on plant-microbe interactions in General Biology II at Northeastern Illinois University (NEIU), a minority-serving institution with a diverse student body. Students grew sugar beets (Beta vulgaris), extracted DNA from the rhizoplane, and used the Department of Energy Systems Biology Knowledgebase (KBase) for bioinformatic analysis to compare microbial relative abundance in fertilized versus unfertilized soil. Over five semesters, the CURE engaged 103 students and leveraged the intuitive KBase platform to make complex sequencing data accessible. Pre/post-course survey data revealed significant increases in student self-assessed research skills, including the ability to explain results and determine the types of data to collect. Furthermore, students reported significant gains in confidence related to experimental design and hypothesis development, alongside a strong increase in familiarity with KBase. Informal faculty feedback indicated high student engagement and appreciation for the real-world connections (e.g. food systems, agriculture, and health). This scalable, low-cost model effectively integrates data science tools into the foundational curriculum, demonstrating a potent strategy for boosting research skills and broadening participation in authentic scientific inquiry among diverse undergraduate students.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2023

Targeted metagenomic assessment reflects critical colonization in battlefield injuries

Current diagnostics and clinical management strategies for combat wounds are based on decisions made by expert clinicians. However, even in the hands of experienced surgeons, wounds from combat injuries can exhibit failed healing and complications related to limitations in the rapid and comprehensive generation of diagnostic information. Previous studies have demonstrated the possible use of genomic sequencing approaches to detect microbial signatures involved in combat casualty care. While effective, whole metagenome sequencing is limited by the depth required to confidently detect all relevant signatures. To address this, we developed a targeted capture sequencing panel to detect microbial signatures relevant to wound healing. These targets include known microbial nosocomial pathogens, wound colonizers, and genes involved in virulence and antimicrobial resistance. A bioinformatics pipeline was built to identify genomic regions of interest and over 8,000 oligonucleotide probes were designed for capture. The panel was synthesized and validated using control reference genomes in human background and on wound-effluent samples from a cohort of combat-injured U.S. service members. Our panel was sensitive against wound-colonizing species, Acinetobacter baumannii and Pseudomonas aeruginosa, and was specific in detecting corresponding virulence and antimicrobial-resistance genes as well as other pathogenic species present in microflora mixtures. Random forest feature permutation confirmed the prevalence of Acinetobacter and Pseudomonas in critically colonized wounds and wounds that failed to heal, respectively. Our results demonstrate the capability of targeted sequencing tools and analysis platforms to profile and deliver information on pathogenic factors influencing wound progression, thereby guiding therapeutic intervention.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2022

$\mathrm{CROPSR}$: an automated platform for complex genome-wide $\mathrm{CRISPR}$ g$\mathrm{RNA}$ design and validation

CRISPR/Cas9 technology has become an important tool to generate targeted, highly specific genome mutations. The technology has great potential for crop improvement, as crop genomes are tailored to optimize specific traits over generations of breeding. Many crops have highly complex and polyploid genomes, particularly those used for bioenergy or bioproducts. The majority of tools currently available for designing and evaluating gRNAs for CRISPR experiments were developed based on mammalian genomes that do not share the characteristics or design criteria for crop genomes. We have developed an open source tool for genome-wide design and evaluation of gRNA sequences for CRISPR experiments, CROPSR. The genome-wide approach provides a significant decrease in the time required to design a CRISPR experiment, including validation through PCR, at the expense of an overhead compute time required once per genome, at the first run. To better cater to the needs of crop geneticists, restrictions imposed by other packages on design and evaluation of gRNA sequences were lifted. A new machine learning model was developed to provide scores while avoiding situations in which the currently available tools sometimes failed to provide guides for repetitive, A/T-rich genomic regions. We show that our gRNA scoring model provides a significant increase in prediction accuracy over existing tools, even in non-crop genomes. CROPSR provides the scientific community with new methods and a new workflow for performing CRISPR/Cas9 knockout experiments. CROPSR reduces the challenges of working in crops, and helps speed gRNA sequence design, evaluation and validation. We hope that the new software will accelerate discovery and reduce the number of failed experiments.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2026

RCSB protein data Bank: Next‐generation advanced search for exploration of experimental structures and computed structure models

Abstract The Protein Data Bank (PDB), established in 1971, is the primary global, open‐access archive for experimentally determined 3D macromolecular structures (proteins, RNA, DNA). The research‐focused RCSB.org web‐portal provides access to these data alongside more than one million machine‐learning‐predicted structure models, greatly expanding the available structural landscape. Rapid growth of both experimental and computational structures has increased the need for powerful yet accessible search tools that serve a broad and diverse scientific community. Herein, we describe a redesigned RCSB Protein Data Bank RCSB.org Advanced Search capability that supports intuitive discovery of 3D structures through a unified interface. This interface integrates annotation‐, sequence‐, and 3D structure‐based searches, embeds an interactive 3D viewer, and incorporates curated biological knowledge, such as catalytic site definitions from Mechanism and Catalytic Site Atlas and ligand‐guided structural motifs, for constructing geometry‐driven queries. A new Chemical Search tool allows definition of chemical queries via an integrated drawing tool or standard identifiers, seamlessly combining them with annotation filters. By allowing query definition directly within spatial and chemical contexts, these search interfaces reduce the need for detailed knowledge of residue numbering, chain identifiers, or external cheminformatics software. This capability enables efficient exploration of structures, chemical diversity, and structure–function relationships across all life domains. The redesigned interfaces can be accessed directly at rcsb.org/search/advanced for Advanced Search and rcsb.org/search/chemical for Chemical Search.

Rose, Yana [Research Collaboratory for Structural ↗
DOE OSTI2022

SBbadger: biochemical reaction networks with definable degree distributions

Abstract Motivation An essential step in developing computational tools for the inference, optimization and simulation of biochemical reaction networks is gauging tool performance against earlier efforts using an appropriate set of benchmarks. General strategies for the assembly of benchmark models include collection from the literature, creation via subnetwork extraction and de novo generation. However, with respect to biochemical reaction networks, these approaches and their associated tools are either poorly suited to generate models that reflect the wide range of properties found in natural biochemical networks or to do so in numbers that enable rigorous statistical analysis. Results In this work, we present SBbadger, a python-based software tool for the generation of synthetic biochemical reaction or metabolic networks with user-defined degree distributions, multiple available kinetic formalisms and a host of other definable properties. SBbadger thus enables the creation of benchmark model sets that reflect properties of biological systems and generate the kinetics and model structures typically targeted by computational analysis and inference software. Here, we detail the computational and algorithmic workflow of SBbadger, demonstrate its performance under various settings, provide sample outputs and compare it to currently available biochemical reaction network generation software. Availability and implementation SBbadger is implemented in Python and is freely available at https://github.com/sys-bio/SBbadger and via PyPI at https://pypi.org/project/SBbadger/. Documentation can be found at https://SBbadger.readthedocs.io. Supplementary information Supplementary data are available at Bioinformatics online.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2023

Energy landscapes from cryo-EM snapshots: a benchmarking study

Abstract Biomolecules undergo continuous conformational motions, a subset of which are functionally relevant. Understanding, and ultimately controlling biomolecular function are predicated on the ability to map continuous conformational motions, and identify the functionally relevant conformational trajectories. For equilibrium and near-equilibrium processes, function proceeds along minimum-energy pathways on one or more energy landscapes, because higher-energy conformations are only weakly occupied. With the growing interest in identifying functional trajectories, the need for reliable mapping of energy landscapes has become paramount. In response, various data-analytical tools for determining structural variability are emerging. A key question concerns the veracity with which each data-analytical tool can extract functionally relevant conformational trajectories from a collection of single-particle cryo-EM snapshots. Using synthetic data as an independently known ground truth, we benchmark the ability of four leading algorithms to determine biomolecular energy landscapes and identify the functionally relevant conformational paths on these landscapes. Such benchmarking is essential for systematic progress toward atomic-level movies of continuous biomolecular function.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2023

Using low-coverage whole genome sequencing (genome skimming) to delineate three introgressed species of buffalofish ( Ictiobus )

Consumption of buffalofish has been sporadically associated with Haff disease-like illnesses involving sudden onset muscle pain and weakness due to skeletal muscle rhabdomyolysis, but determination of precisely which species are associated with these illnesses has been impeded by a lack of species-specific DNA-based markers. Here, three closely related species of buffalofish native to the Mississippi River Basin (Ictiobus bubalus, Ictiobus cyprinellus and Ictiobus niger) that have previously proven genetically indistinguishable using both mitochondrial and nuclear single-locus sequencing were reliably discriminated using low-coverage whole genome sequencing (‘genome skimming’). Using 44 specimens representing the three species collected from the mid/upper (Missouri) and lower (Louisiana) regions of the species’ native ranges, the SISRS (Site Identification from Short Read Sequences) bioinformatics pipeline was adapted to (1) identify over 620Mbp of putatively homologous nuclear sequence data and (2) isolate over 140,000 single-nucleotide polymorphisms (SNPs) that supported accurate species delimitation, all without the use of a reference genome or annotation data. These sites were used to classify Ictiobus spp. samples with genome-skim data, along with a larger set (n = 67) where ultraconserved elements (UCEs) were sequenced. Analyses of whole mitochondrial data revealed more limited signal. Nearly all samples matched their purported species based on morphologic identification, but two Missouri samples morphologically identified as I. niger grouped with samples of I. bubalus, albeit with significant enrichment of I. niger SNPs. To our knowledge this is the first report of a DNA-based tool to reliably discriminate these three morphologically distinct species.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2021

EcoPLOT: dynamic analysis of biogeochemical data

Motivation: We have created EcoPLOT (parameterized linkage of omics-driven technologies), a web-app for the dynamic, interactive analysis of biogeochemical datasets that combines state-of-the-art analysis tools to statistically and graphically explore environmental, geochemical and microbiome datasets. Using the iterative random forest, a machine learning algorithm, EcoPLOT allows for the de novo discovery of drivers which exhibit significant impact on plant, microbial or soil dynamics. Availability and implementation: EcoPLOT is built entirely within the R language. It can be accessed through any system where R is installed, including Windows, Mac and most Linux systems. EcoPLOT is free to use and can be accessed at https://github.com/cdsanchez18/EcoPLOT.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2024

HDBind: encoding of molecular structure with hyperdimensional binary representations

Traditional methods for identifying “hit” molecules from a large collection of potential drug-like candidates rely on biophysical theory to compute approximations to the Gibbs free energy of the binding interaction between the drug and its protein target. These approaches have a significant limitation in that they require exceptional computing capabilities for even relatively small collections of molecules. Increasingly large and complex state-of-the-art deep learning approaches have gained popularity with the promise to improve the productivity of drug design, notorious for its numerous failures. However, as deep learning models increase in their size and complexity, their acceleration at the hardware level becomes more challenging. Hyperdimensional Computing (HDC) has recently gained attention in the computer hardware community due to its algorithmic simplicity relative to deep learning approaches. The HDC learning paradigm, which represents data with high-dimension binary vectors, allows the use of low-precision binary vector arithmetic to create models of the data that can be learned without the need for the gradient-based optimization required in many conventional machine learning and deep learning methods. This algorithmic simplicity allows for acceleration in hardware that has been previously demonstrated in a range of application areas (computer vision, bioinformatics, mass spectrometery, remote sensing, edge devices, etc.). To the best of our knowledge, our work is the first to consider HDC for the task of fast and efficient screening of modern drug-like compound libraries. We also propose the first HDC graph-based encoding methods for molecular data, demonstrating consistent and substantial improvement over previous work. We compare our approaches to alternative approaches on the well-studied MoleculeNet dataset and the recently proposed LIT-PCBA dataset derived from high quality PubChem assays. We demonstrate our methods on multiple target hardware platforms, including Graphics Processing Units (GPUs) and Field Programmable Gate Arrays (FPGAs), showing at least an order of magnitude improvement in energy efficiency versus even our smallest neural network baseline model with a single hidden layer. Our work thus motivates further investigation into molecular representation learning to develop ultra-efficient pre-screening tools. We make our code publicly available at https://github.com/LLNL/hdbind.

59 BASIC BIOLOGICAL SCIENCES↗