Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “mutation”

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 523 records · Page 29
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

BBO-10203 inhibits tumor growth without inducing hyperglycemia by blocking RAS-PI3Kα interaction

BBO-10203 is an orally available drug that covalently and specifically binds to the rat sarcoma (RAS)–binding domain of phosphoinositide 3-kinase α (PI3Kα), preventing its activation by HRAS, NRAS, and KRAS. Here, it inhibited PI3Kα activation in tumors with oncogenic mutations in KRAS or PIK3CA and in tumors with human epidermal growth factor receptor 2 (HER2) amplification or overexpression. In preclinical models, BBO-10203 caused significant tumor growth inhibition across multiple tumor types and showed enhanced efficacy in combination with inhibitors of cyclin-dependent kinase 4/6 (CDK4/6), estrogen receptor (ER), HER2, and KRAS-G12C mutant, including in tumors harboring mutations in Kelch-like ECH-associated protein 1 (KEAP1) and serine/threonine kinase 11 (STK11). Notably, these antitumor effects occurred without inducing hyperglycemia, because insulin signaling does not depend on RAS-mediated PI3Kα activation to promote glucose uptake.

Simanshu, Dhirendra K. [Frederick National Laborat↗
DOE OSTI2025

Secure biosystems design in Saccharomyces cerevisiae establishes effective biocontainment strategies and mechanisms of escape

The widespread application of recombinant DNA and synthetic biology approaches for microbial metabolic engineering pursuits has motivated the development of biocontainment strategies, targeting safe and secure deployment of genetically modified microorganisms (GMMs). However, the design rules and mechanistic drivers governing biocontainment efficacy, as well as impacts of biocontainment upon microbial fitness, remain to be comprehensively evaluated, hindering predictive design and application of these strategies. We have developed a platform for high-resolution analysis of a transactivated kill switch in laboratory and industrial strains of Saccharomyces cerevisiae to assess modes of biocontainment escape and establish design rules for development of kill switch systems in diverse microbes. A camphor-regulated, RelE toxin system was systematically deployed to assess the impacts of differential kill switch copy number and ploidy in laboratory vs industrial strains. CRISPR-mediated integration of the biocontainment system at various loci revealed rapid escape events driven, in part, by mutations to both the Cam-transactivator (cam-TA) and RelE toxin. Genetic engineering enabled recapitulation of escape phenotypes, confirming mechanisms of escape and establishing structure-function relationships in the cam-TA system. Interestingly, genomic resequencing of escape mutants also revealed a series of off-target mutations, implicating additional modes of kill switch escape. Multi-copy integration of the kill switch system mitigated these effects by orders of magnitude, without compromising the biosynthetic capacity of the microbes, but proved insufficient to establish sustained biocontainment. The resultant data define a series of key design rules for next-generation biocontainment strategies and add to a growing foundational knowledge base targeting establishment of secure biosystems designs.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2024

Genetic and enzymatic characterization of Amy13E from Cellvibrio japonicus reclassifies it as a cyclodextrinase also capable of α-diglucoside degradation

ABSTRACT Cyclodextrinases are carbohydrate-active enzymes involved in the linearization of circular amylose oligosaccharides. Primarily thought to function as part of starch metabolism, there have been previous reports of bacterial cyclodextrinases also having additional enzymatic activities on linear malto-oligosaccharides. This substrate class also includes environmentally rare α-diglucosides such as kojibiose (α−1,2), nigerose (α−1,3), and isomaltose (α−1,6), all of which have valuable properties as prebiotics or low-glycemic index sweeteners. Previous genome sequencing of threeCellvibrio japonicusstrains adapted to utilize these α-diglucosides identified multiple, but uncharacterized, mutations in each strain. One of the mutations identified was in theamy13Egene, which was annotated to encode a neopullulanase. In this report, we functionally characterized this gene and determined that it in fact encodes a cyclodextrinase with additional activities on α-diglucosides. Deletion analysis ofamy13Efound that this gene was essential for kojibiose and isomaltose metabolism inC. japonicus. Interestingly, a Δamy13Emutant was not deficient for cyclodextrin or pullulan utilization inC. japonicus; however, heterologous expression of the gene inE. coliwas sufficient for cyclodextrin-dependent growth. Biochemical analyses found thatCjAmy13E cleaved multiple substrates but preferred cyclodextrins and maltose, but had no activity on pullulan. Our characterization of theCjAmy13E cyclodextrinase is useful for refining functional enzyme predictions in related bacteria and for engineering enzymes for biotechnology or biomedical applications. IMPORTANCE Understanding the bacterial metabolism of cyclodextrins and rare α-diglucosides is increasingly important, as these sugars are becoming prevalent in the foods, supplements, and medicines humans consume that subsequently feed the human gut microbiome. Our analysis of a cyclomaltodextrinase with an expanded substrate range is significant because it broadens the potential applications of the GH13 family of carbohydrate active enzymes (CAZymes) in biotechnology and biomedicine. Specifically, this study provides a workflow for the discovery and characterization of novel activities in bacteria that possess a high number of CAZymes that otherwise would be missed due to complications with functional redundancy. Furthermore, this study provides a model from which predictions can be made why certain bacteria in crowded niches are able to robustly utilize rare carbon sources, possibly to gain a competitive growth advantage.

Biotechnology & Applied Microbiology↗
DOE OSTI2020

Trehalose Degradation by Cellvibrio japonicus Exhibits No Functional Redundancy and Is Solely Dependent on the Tre37A Enzyme

The α-diglucoside trehalose has historically been known as a component of the bacterial stress response, though it more recently has been studied for its relevance in human gut health and biotechnology development. The utilization of trehalose as a nutrient source by bacteria relies on carbohydrate-active enzymes, specifically those of the glycoside hydrolase family 37 (GH37), to degrade the disaccharide into substituent glucose moieties for entry into metabolism. Environmental bacteria using oligosaccharides for nutrients often possess multiple carbohydrate-active enzymes predicted to have the same biochemical activity and therefore are thought to be functionally redundant. In this study, we characterized trehalose degradation by the biotechnologically important saprophytic bacterium Cellvibrio japonicus This bacterium possesses two predicted α-α-trehalase genes, tre37A and tre37B, and our investigation using mutational analysis found that only the former is essential for trehalose utilization by C. japonicus Heterologous expression experiments found that only the expression of the C. japonicus tre37A gene in an Escherichia coli treA mutant strain allowed for full utilization of trehalose. Biochemical characterization of C. japonicus GH37 activity determined that the tre37A gene product is solely responsible for cleaving trehalose and is an acidic α-α-trehalase. Bioinformatic and mutational analyses indicate that Tre37A directly cleaves trehalose to glucose in the periplasm, as C. japonicus does not possess a phosphotransferase system. This study facilitates the development of a comprehensive metabolic model for α-linked disaccharides in C. japonicus and more broadly expands our understanding of the strategies that saprophytic bacteria employ to capture diverse carbohydrates from the environment. The metabolism of trehalose is becoming increasingly important due to the inclusion of this α-diglucoside in a number of foods and its prevalence in the environment. Bacteria able to utilize trehalose in the human gut possess a competitive advantage, as do saprophytic microbes in terrestrial environments. While the biochemical mechanism of trehalose degradation is well understood, what is less clear is how bacteria acquire this metabolite from the environment. The significance of this report is that by using the model saprophyte Cellvibrio japonicus, we were able to functionally characterize the two predicted trehalase enzymes that the bacterium possesses and determined that the two enzymes are not equivalent and are not functionally redundant. Here, the results and approaches used to understand the complex physiology of α-diglucoside metabolism from this study can be applied broadly to other polysaccharide-degrading bacteria.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2025

Polyyne production is regulated by the transcriptional regulators PgnC and GacA in Pseudomonas protegens Pf-5

ABSTRACT Polyynes produced by bacteria have promising applications in agriculture and medicine due to their potent antimicrobial activities. Polyyne biosynthetic genes have been identified inPseudomonasandBurkholderia. However, the molecular mechanisms underlying the regulation of polyyne biosynthesis remain largely unknown. In this study, we used a soil bacteriumPseudomonas protegensPf-5, which was recently reported to produce polyyne called protegenin, as a model to investigate the regulation of bacterial polyyne production. Our results show that Pf-5 controls polyyne production at both the pathway-specific level and a higher global level. Mutation ofpgnC, a transcriptional regulatory gene located in the polyyne biosynthetic gene cluster, abolished polyyne production. Gene expression analysis revealed that PgnC directly activates the promoter of polyyne biosynthetic genes. The production of polyyne also requires a global regulator GacA. Mutation ofgacAdecreased the translation of PgnC, which is consistent with the result thatpgnCleader mRNA bound directly to RsmE, an RNA-binding protein negatively regulated by GacA. These results suggest that GacA induces the expression of the PgnC regulator, which in turn activates polyyne biosynthesis. Additionally, the polyyne-producing strain of Pf-5, but not the polyyne-nonproducing strain, could inhibit a broad spectrum of bacteria including both Gram-negative and Gram-positive bacteria. IMPORTANCE Antimicrobial metabolites produced by bacteria are widely used in agriculture and medicine to control plant, animal, and human pathogens. Although bacteria-derived polyynes have been identified as potent antimicrobials for decades, the molecular mechanisms by which bacteria regulate polyyne biosynthesis remain understudied. In this study, we found that polyyne biosynthesis is directly activated by a pathway-specific regulator PgnC, which is induced by a global regulator GacA through the RNA-binding protein RsmE inPseudomonas protegens. To our knowledge, this work is the first comprehensive study of the regulatory mechanisms of bacterial polyyne biosynthesis at both pathway-specific level and global level. The discovered molecular mechanisms can help us optimize polyyne production for agricultural or medical applications.

Biotechnology & Applied Microbiology↗
DOE OSTI2020

Combining Random Mutagenesis and Metabolic Engineering for Enhanced Tryptophan Production in Synechocystis sp. Strain PCC 6803

Tryptophan (Trp) is an essential aromatic amino acid that has value as an animal feed supplement, as the amount found in plant-based sources is insufficient. An alternative to production by engineered microbial fermentation is to have tryptophan biosynthesized by a photosynthetic microorganism that could replace or supplement both the plant and industrially used microbes. We selected Synechocystis sp. strain PCC 6803, a model cyanobacterium, as the host and studied metabolic engineering and random mutagenesis approaches. Previous work on engineering heterotrophic microbes for improved Trp titers has targeted allosteric feedback regulation in enzymes 3-deoxy-d-arabinoheptulosonate 7-phosphate synthase (DAHPS) and anthranilate synthase (AS) as major bottlenecks in the shikimate pathway. In this work, the genes encoding feedback-resistant enzymes from Escherichia coli, aroG fbr and trpE fbr , were overexpressed in the host wild-type (WT) strain. Separately, the WT strain was subjected to random mutagenesis and selection using an amino acid analog to isolate tryptophan-overproducing strains. The randomly mutagenized strains were sequenced in order to identify the mutations that resulted in the desirable phenotypes. Interestingly, the tryptophan overproducers had mutations in the gene encoding chorismate mutase (CM), which catalyzes the conversion of chorismate to prephenate. The best tryptophan overproducer from random mutagenesis was selected as a host for metabolic engineering where aroG fbr and trpE fbr were overexpressed. The best strain developed produced 212 ± 23 mg/liter of tryptophan after 10 days of photoautotrophic growth under 3% (vol/vol) CO2. We demonstrated that a combination of random mutagenesis and metabolic engineering was superior to either individual approach.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2020

Type IV-like pili facilitate transformation in naturally competent archaea

Naturally competent organisms are capable of DNA uptake directly from the environment through the process of transformation. Despite the importance of transformation to microbial evolution, DNA uptake remains poorly characterized outside the bacterial domain. Here, we identify the pilus as a necessary component of the transformation machinery in archaea. We describe two naturally competent organisms: Methanococcus maripaludis and Methanoculleus thermophilus. In M. maripaludis, replicative vectors were transferred with an average efficiency of 2.4 x 10 3 transformants μg -1 DNA. In M. thermophilus, integrative vectors were transferred with an average efficiency of 2.7 x 10 3 transformants μg -1 DNA. Additionally, natural transformation of M. thermophilus could be used to introduce chromosomal mutations. To our knowledge, this is the first demonstration of a method to introduce targeted mutations in a member of the order Methanomicrobiales. For both organisms, mutants lacking structural components of the type IV-like pilus filament were defective for DNA uptake, demonstrating the importance of pili for natural transformation. Interestingly, competence could be induced in a non-competent strain of M. maripaludis by expressing pilin genes from a replicative vector. Furthermore, these results expand the known natural competence pili to include examples from the archaeal domain and highlight the importance of pili for DNA uptake in diverse microbial organisms.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2022

Structure-Based Design and Antigenic Validation of Respiratory Syncytial Virus G Immunogens

Respiratory syncytial virus (RSV) is a leading cause of severe lower respiratory tract disease of children, the elderly, and immunocompromised individuals. Currently, there are no FDA-approved RSV vaccines. The RSV G glycoprotein is used for viral attachment to host cells and impairment of host immunity by interacting with the human chemokine receptor CX3CR1. Antibodies that disrupt this interaction are protective against infection and disease. Nevertheless, development of an RSV G vaccine antigen has been hindered by its low immunogenicity and safety concerns. A previous study described three engineered RSV G proteins containing single-point mutations that induce higher levels of IgG antibodies and have improved safety profiles compared to wild-type RSV G. However, it is unclear if the mutations affect RSV G protein folding and display of its conformational epitopes. In this study, we show that the RSV G S177Q protein retains high-affinity binding to protective human and mouse monoclonal antibodies and has equal reactivity as wild-type RSV G protein to human reference immunoglobulin to RSV. Additionally, we determined the high-resolution crystal structure of RSV G S177Q protein in complex with the anti-RSV G antibody 3G12, further validating its antigenic structure. These studies show for the first time that an engineered RSV G protein with increased immunogenicity and safety retains conformational epitopes to high-affinity protective antibodies, supporting its further development as an RSV vaccine immunogen. Respiratory syncytial virus (RSV) causes severe lower respiratory diseases of children, the elderly, and immunocompromised populations. There currently are no FDA-approved RSV vaccines. Most vaccine development efforts have focused on the RSV F protein, and the field has generally overlooked the receptor-binding antigen RSV G due to its poor immunogenicity and safety concerns. However, single-point mutant RSV G proteins have been previously identified that have increased immunogenicity and safety. In this study, we investigate the antibody reactivities of three known RSV G mutant proteins. We show that one mutant RSV G protein retains high-affinity binding to protective monoclonal antibodies, is equally recognized by anti-RSV antibodies in human sera, and forms the same three-dimensional structure as the wild-type RSV G protein. Our study validates the structure-guided design of the RSV G protein as an RSV vaccine antigen.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2020

Estimating the Timing of Early Simian-Human Immunodeficiency Virus Infections: a Comparison between Poisson Fitter and BEAST

Many HIV prevention strategies are currently under consideration where it is highly informative to know the study participants’ times of infection. These can be estimated using viral sequence data sampled early in infection. However, there are several scenarios that, if not addressed, can skew timing estimates. These include multiple transmitted/founder (TF) viruses, APOBEC (apolipoprotein B mRNA editing enzyme, catalytic polypeptide-like)-mediated mutational enrichment, and recombination. Here, we suggest a pipeline to identify these problems and resolve the biases that they introduce. We then compare two modeling strategies to obtain timing estimates from sequence data. The first, Poisson Fitter (PF), is based on a Poisson model of random accumulation of mutations relative to the TF virus (or viruses) that established the infection. The second uses a coalescence-based phylogenetic strategy as implemented in BEAST. The comparison is based on timing predictions using plasma viral RNA (cDNA) sequence data from 28 simian-human immunodeficiency virus (SHIV)-infected animals for which the exact day of infection is known. In this particular setting, based on nucleotide sequences from samples obtained in early infection, the Poisson method yielded more accurate, more precise, and unbiased estimates for the time of infection than did the explored implementations of BEAST. The inference of the time of infection is a critical parameter in testing the efficacy of clinical interventions in protecting against HIV-1 infection. For example, in clinical trials evaluating the efficacy of passively delivered antibodies (Abs) for preventing infections, accurate time of infection data are essential for discerning levels of the Abs required to confer protection, given the natural Ab decay rate in the human body. In such trials, genetic sequences from early in the infection are regularly sampled from study participants, generally prior to immune selection, when the viral population is still expanding and genetic diversity is low. In this particular setting of early viral growth, the Poisson method is superior to the alternative approach based on coalescent methods. This approach can also be applied in human vaccine trials, where accurate estimates of infection times help ascertain if vaccine-elicited immune protection wanes over time.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2020

Effects of Genetic and Physiological Divergence on the Evolution of a Sulfate-Reducing Bacterium under Conditions of Elevated Temperature

ABSTRACT Adaptation via natural selection is an important driver of evolution, and repeatable adaptations of replicate populations, under conditions of a constant environment, have been extensively reported. However, isolated groups of populations in nature tend to harbor both genetic and physiological divergence due to multiple selective pressures that they have encountered. How this divergence affects adaptation of these populations to a new common environment remains unclear. To determine the impact of prior genetic and physiological divergence in shaping adaptive evolution to accommodate a new common environment, an experimental evolution study with the sulfate-reducing bacterium Desulfovibrio vulgaris Hildenborough (DvH) was conducted. Two groups of replicate populations with genetic and physiological divergence, derived from a previous evolution study, were propagated in an elevated-temperature environment for 1,000 generations. Ancestor populations without prior experimental evolution were also propagated in the same environment as a control. After 1,000 generations, all the populations had increased growth rates and all but one had greater fitness in the new environment than the ancestor population. Moreover, improvements in growth rate were moderately affected by the divergence in the starting populations, while changes in fitness were not significantly affected. The mutations acquired at the gene level in each group of populations were quite different, indicating that the observed phenotypic changes were achieved by evolutionary responses that differed between the groups. Overall, our work demonstrated that the initial differences in fitness between the starting populations were eliminated by adaptation and that phenotypic convergence was achieved by acquisition of mutations in different genes. IMPORTANCE Improving our understanding of how previous adaptation influences evolution has been a long-standing goal in evolutionary biology. Natural selection tends to drive populations to find similar adaptive solutions for the same selective conditions. However, variations in historical environments can lead to both physiological and genetic divergence that can make evolution unpredictable. Here, we assessed the influence of divergence on the evolution of a model sulfate-reducing bacterium, Desulfovibrio vulgaris Hildenborough, in response to elevated temperature and found a significant effect at the genetic but not the phenotypic level. Understanding how these influences drive evolution will allow us to better predict how bacteria will adapt to various ecological constraints.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2024

Structure and in vivo psoralen DNA crosslink repair activity of mycobacterial Nei2

Mycobacterium smegmatis Nei2 is a monomeric enzyme with AP β-lyase activity on single-stranded DNA. Expression of Nei2, and its operonic neighbor Lhr (a tetrameric 3'-to-5' helicase), is induced in mycobacteria exposed to DNA damaging agents. Here, we find that nei2 deletion sensitizes M. smegmatis to killing by DNA inter-strand crosslinker trimethylpsoralen but not to crosslinkers mitomycin C and cisplatin. By contrast, deletion of lhr sensitizes to killing by all three crosslinking agents. We report a 1.45 Å crystal structure of recombinant Nei2, which is composed of N and C terminal lobes flanking a central groove suitable for DNA binding. The C lobe includes a tetracysteine zinc complex. Mutational analysis identifies the N-terminal proline residue (Pro2 of the ORF) and Lys51, but not Glu3, as essential for AP lyase activity. We find that Nei2 has 5-hydroxyuracil glycosylase activity on single-stranded DNA that is effaced by alanine mutations of Glu3 and Lys51 but not Pro2. Testing complementation of psoralen sensitivity by expression of wild-type and mutant nei2 alleles in Δnei2 cells established that AP lyase activity is neither sufficient nor essential for crosslink repair. By contrast, complementation of psoralen sensitivity of Δlhr cells by mutant lhr alleles depended on Lhr’s ATPase/helicase activities and its tetrameric quaternary structure. The lhr–nei2 operon comprises a unique bacterial system to rectify inter-strand crosslinks.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2021

Highly Basic Clusters in the Herpes Simplex Virus 1 Nuclear Egress Complex Drive Membrane Budding by Inducing Lipid Ordering

During replication of herpesviruses, capsids escape from the nucleus into the cytoplasm by budding at the inner nuclear membrane. This unusual process is mediated by the viral nuclear egress complex (NEC) that deforms the membrane around the capsid by oligomerizing into a hexagonal, membrane-bound scaffold. Here, we found that highly basic membrane-proximal regions (MPRs) of the NEC alter lipid order by inserting into the lipid headgroups and promote negative Gaussian curvature. We also find that the electrostatic interactions between the MPRs and the membranes are essential for membrane deformation. One of the MPRs is phosphorylated by a viral kinase during infection, and the corresponding phosphomimicking mutations block capsid nuclear egress. We show that the same phosphomimicking mutations disrupt the NEC-membrane interactions and inhibit NEC-mediated budding in vitro, providing a biophysical explanation for the in vivo phenomenon. Our data suggest that the NEC generates negative membrane curvature by both lipid ordering and protein scaffolding and that phosphorylation acts as an off switch that inhibits the membrane-budding activity of the NEC to prevent capsid-less budding.

membrane curvature↗
DOE OSTI2023

Influenza A Virus Multicycle Replication Yields Comparable Viral Population Emergence in Human Respiratory and Ocular Cell Types

While primarily considered a respiratory pathogen, influenza A virus (IAV) is nonetheless capable of spreading to, and replicating in, numerous extrapulmonary tissues in humans. However, within-host assessments of genetic diversity during multicycle replication have been largely limited to respiratory tract tissues and specimens. As selective pressures can vary greatly between anatomical sites, there is a need to examine how measures of viral diversity may vary between influenza viruses exhibiting different tropisms in humans, as well as following influenza virus infection of cells derived from different organ systems. Here, we employed human primary tissue constructs emulative of the human airway or corneal surface, and we infected both with a panel of human- and avian-origin IAV, inclusive of H1 and H3 subtype human viruses and highly pathogenic H5 and H7 subtype viruses, which are associated with both respiratory disease and conjunctivitis following human infection. While both cell types supported productive replication of all viruses, airway-derived tissue constructs elicited greater induction of genes associated with antiviral responses than did corneal-derived constructs. We used next-generation sequencing to examine viral mutations and population diversity, utilizing several metrics. With few exceptions, generally comparable measures of viral diversity and mutational frequency were detected following homologous virus infection of both respiratory-origin and ocular-origin tissue constructs. Expansion of within-host assessments of genetic diversity to include IAV with atypical clinical presentations in humans or in extrapulmonary cell types can provide greater insight into understanding those features most prone to modulation in the context of viral tropism.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2025

A minimal SufB 2 C 2 complex functions as a [4Fe-4S] cluster scaffold in methanogenic archaea

Iron-sulfur clusters are essential cofactors in all domains of life, yet their biogenesis in obligately anaerobic archaea remains poorly understood. Here, we characterized the minimal two-protein SUF system in methanogenic archaea, composed solely of SufB and SufC. Using Methanococcus maripaludis as a model, we demonstrate that the SUF proteins from its native host form a stable SufB 2 C 2 heterotetramer that binds a [4Fe-4S] cluster via three conserved cysteines in SufC. Mutations of conserved cysteine and histidine residues of SufB do not impair cluster binding. The complex interacts with the SAM-containing methanogenesis marker protein 10 (MmpX), suggesting direct Fe-S cluster transfer from SufB 2 C 2 to target proteins. Mutational analysis of Methanothermococcus thermolithotrophicus proteins confirmed that SufC is the primary cluster-binding component, while SufB enhances ATPase and cluster transfer activities. Evolutionary comparisons suggest that this two-protein SUF system represents an ancestral form of Fe-S cluster biogenesis.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2024

Efficacy of using a dynamic length representation vs. a fixed-length for neuroarchitecture search

Deep learning neuroarchitecture and hyperparameter search are important in finding the best configuration that maximizes learned model accuracy. However, the number of types of layers, their associated hyperparameters, and the myriad of ways to connect layers poses a significant computational challenge in discovering ideal model configurations. Here, we assess two different approaches for neuroarchitecture search for a LeNet style neural network, one that uses a fixed-length approach where there is a preset number of possible layers that can be toggled on or off via mutation, and a variable-length approach where layers can be freely added or removed via special mutation operators. We found that the variable-length implementation trained better models while discovering unusual layer configurations worth further exploration.

Coletti, Mark↗
DOE OSTI2023

Mass Spectrometry–Based Proteogenomics: New Therapeutic Opportunities for Precision Medicine

Proteogenomics refers to the integration of comprehensive genomic, transcriptomic, and proteomic measurements from the same samples with the goal of fully understanding the regulatory processes converting genotypes to phenotypes, often with an emphasis on gaining a deeper understanding of disease processes. Although specific genetic mutations have long been known to drive the development of multiple cancers, gene mutations alone do not always predict prognosis or response to targeted therapy. The benefit of proteogenomics research is that information obtained from proteins and their corresponding pathways provides insight into therapeutic targets that can complement genomic information by providing an additional dimension regarding the underlying mechanisms and pathophysiology of tumors. This review describes the novel insights into tumor biology and drug resistance derived from proteogenomic analysis while highlighting the clinical potential of proteogenomic observations and advances in technique and analysis tools.

60 APPLIED LIFE SCIENCES↗
DOE OSTI2021

HUNTRESS: a fast heuristic for reconstructing phylogenetic trees of tumor evolution (HUNTRESS) v0.1

We introduce HUNTRESS (Histogrammed UNion Tree REconStruction heuriStic), a computational method for tumor phylogeny reconstruction from noisy genotype matrices derived from single-cell sequencing data, whose running time is linear with the number of cells and quadratic with the number of mutations. Provided that the input genotype matrix includes no false positives, each cellular subpopulation is at least a user defined fraction of the total number of cells, and the number of cells are much bigger than the number of mutations considered, HUNTRESS computes the ground truth tumor phylogeny with high probability. On simulated data HUNTRESS is faster than available alternatives with comparable or better accuracy. Additionally, the phylogenies reconstructed by HUNTRESS on two single-cell sequencing data sets agree with the best known evolutionary scenarios for the associated tumors.

Buluc, Aydin↗
DOE OSTI2022

FutureTense

Protective vaccines and reliable diagnostics are essential tools for controlling viral diseases. However, the efficacy of these tools can be diminished by mutations in viral genomes. The delay between the emergence of new viral strains and the redesign of vaccines and diagnostics allows for continued viral transmission. Is it possible to address this challenge by computationally predicting viral genome sequence evolution? Can we “future-proof” vaccines and diagnostics by targeting both current and anticipated future sequence variants? While predicting viral evolution is still an unsolved, “grand challenge” problem in biology, the large, and rapidly growing, number of SARS-CoV-2 genome sequences provide an opportunity to quantify the ability of machine learning to predict viral genome sequence evolution. Towards this end, we have developed a simple computational model for predicting viral evolution at the level of individual nucleotides. The key metric for quantifying the per-base, prediction accuracy for viral evolution is the Mann-Whitney U statistic (or, equivalently, the area under the receiver operator curve). Since the Mann-Whitney U statistic is not a differentiable function, existing deep leaning packages (like Pytorch and Keras/TensorFlow) are not useful, as they require that the accuracy metric/objective function be analytically differentiable with respect to the model parameters. To overcome this challenge, we have implemented custom software, “FutureTense”, that can train a machine learning model by maximizing the non-differentiable Mann-Whitney U statistic. This software trains a machine learning model by exploring along the direction of the discrete gradient of the Mann-Whitney U statistic in the model parameter space. Parallel computing and genome sequence-specific optimizations are used to accelerate model training. The resulting machine learning model learns the observed high C->U mutation rates in the SARS-CoV-2 genome (which are potentially induced by host defenses) and provides prediction accuracies that are significantly better than one would expect from random chance. While predicting viral evolution is still quite far from a solved problem, the surprising performance of this simple model gives hope that the accuracy of predicting viral genome evolution can be further increased by more sophisticated approaches.

Gans, Jason↗