Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “encoding”

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 1,045 records · Page 58
All sourcesNASA NTRSDOE OSTI
DOE OSTI2020

P finder: genomic and metagenomic annotation of RNase P RNA gene (rnpB)

Abstract Background The rnpB gene encodes for an essential catalytic RNA (RNase P). Like other essential RNAs, RNase P’s sequence is highly variable. However, unlike other essential RNAs (i.e. tRNA, 16 S, 6 S,...) its structure is also variable with at least 5 distinct structure types observed in prokaryotes. This structural variability makes it labor intensive and challenging to create and maintain covariance models for the detection of RNase P RNA in genomic and metagenomic sequences. The lack of a facile and rapid annotation algorithm has led to the rnpB gene being the most grossly under annotated essential gene in completed prokaryotic genomes with only a 24% annotation rate. Here we describe the coupling of the largest RNase P RNA database with the local alignment scoring algorithm to create the most sensitive and rapid prokaryote rnpB gene identification and annotation algorithm to date. Results Of the 2772 completed microbial genomes downloaded from GenBank only 665 genomes had an annotated rnpB gene. We applied P Finder to these genomes and were able to identify 2733 or nearly 99% of the 2772 microbial genomes examined. From these results four new rnpB genes that encode the minimal T-type P RNase P RNAs were identified computationally for the first time. In addition, only the second C-type RNase P RNA was identified in Sphaerobacter thermophilus . Of special note, no RNase P RNAs were detected in several obligate endosymbionts of sap sucking insects suggesting a novel evolutionary adaptation. Conclusions The coupling of the largest RNase P RNA database and associated structure class identification with the P Finder algorithm is both sensitive and rapid, yielding high quality results to aid researchers annotating either genomic or metagenomic data. It is the only algorithm to date that can identify challenging RNAse P classes such as C-type and the minimal T-type RNase P RNAs. P Finder is written in C# and has a user-friendly GUI that can run on multiple 64-bit windows platforms (Windows Vista/7/8/10). P Finder is free available for download at https://github.com/JChristopherEllis/P-Finder as well as a small sample RNase P RNA file for testing.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2022

Method for quantification of porcine type I interferon activity using luminescence, by direct and indirect means

Abstract Background Type I interferons are widely used in research applications and as biotherapeutics. Current assays used to measure interferon concentrations, such as plaque reduction assays and ELISA, are expensive, technically challenging, and may take days to provide results. We sought to develop a robust and rapid assay to determine interferon concentrations produced from transiently transfected cell cultures. Method Indirect quantification of recombinant interferon was evaluated using a novel bi-cistronic construct encoding the Foot-and-mouth disease virus 2A translational interrupter sequence to yield equimolar expression of Gaussia princeps luciferase and porcine interferon α. Direct quantification was evaluated by expression of a novel fusion protein comprised of Gaussia princeps luciferase and porcine type I interferon. Plasmids encoding constructs are transiently transfected into cell cultures and supernatant harvested for testing of luminescence, ELISA determined concentration, and anti-viral activity against vesicular stomatitis virus. Results Bi-cistronic constructs, utilized for indirect quantification, demonstrate both luciferase activity and anti-viral activity. Fusion proteins, utilized for direct quantification, retained secretion and luminescence however only the interferon α fusion protein had antiviral activity comparable to wildtype porcine interferon α. A strong linear correlation was observed between dilution and luminescence for all compounds over a dynamic range of concentrations. Conclusion The correlation of antiviral and luciferase activities demonstrated the utility of this approach, both direct and indirect, to rapidly determine recombinant interferon concentrations. Concentration can be determined over a more dynamic concentration range than available ELISA based assays using this methodology.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2026

Multitask graph neural networks for elastoplastic response prediction in dual-phase polycrystals

Microstructure-sensitive prediction of elastoplastic response remains a recurring bottleneck in multiscale damage and fatigue modeling, where large ensembles of statistically distinct polycrystals are required to quantify variability and extreme-value behavior. In this work, we develop a multitask graph neural network (GNN) surrogate that maps dual-phase ferrite–martensite polycrystal microstructures to Statistical Volume Element (SVE)-level elastoplastic Quantities of Interest (QoIs). Each SVE is represented as a grain-adjacency graph, with node features encoding phase, geometry, and crystallographic orientation, and edge features encoding relative misorientation. A message-passing graph convolution generates node embeddings, which are pooled into a graph representation and passed to a multitask regression head that jointly predicts 10 scalar QoIs and vector-valued stress–strain responses in orthogonal loading directions across multiple martensite volume fractions and SVE sizes. Results show high accuracy for scalar QoIs and strong agreement for full stress–strain trajectories, with population envelopes reproducing both median behavior and finite-SVE variability across compositions and partition scales. A unified model trained on pooled volume-fraction data preserves most within-regime accuracy relative to regime-specific models while also capturing the broader cross-regime variation reflected in the pooled test set. Distributional comparisons further demonstrate that the surrogate preserves heterogeneity under SVE partitioning, enabling statistically consistent block-wise random-field construction for mesoscale analyses. Overall, the proposed grain-graph surrogate provides a practical pathway to accelerate ensemble-based studies of SVE-level constitutive variability in dual-phase polycrystals.

Crystal plasticity↗
DOE OSTI2026

Mitigating Algorithmic Bias in Cancer Site Classification Models

Purpose Integrating artificial intelligence in cancer diagnostics has improved tumor classification beyond rule-based systems. Despite these advancements, these models may still encode demographic biases. We conducted a large-scale, applied bias-probing study of a deep learning–based cancer site classifier to quantify race information encoded in document embeddings. We then evaluated how performance changes when race-correlated embedding dimensions are removed in a post-training sensitivity analysis. Methods The cancer site classifier was trained using 3.5 million electronic cancer pathology reports from six of the National Cancer Institute's SEER registries. We trained a hierarchical self-attention network to generate 400-dimensional document embeddings. These embeddings were used to train two downstream, gradient-boosted decision tree classifiers: one to classify the cancer sites and another to predict racial categories. We identified overlapping features by intersecting the top 50 feature-importance rankings from the site and race models and computed their cumulative feature importance in each model. As a post hoc sensitivity analysis, we progressively pruned these overlapping dimensions, retrained the site model, and compared overall macro-F1 and accuracy, race-stratified macro-F1, and group fairness metrics on the basis of demographic parity and equalized odds before and after pruning. Results The analysis revealed minimal feature overlap between the cancer site and race prediction models, and the cumulative importance scores indicated a negligible influence of racial information on clinical predictions. Post-training pruning of overlapping features did not compromise the models' diagnostic accuracy, with a 0.07% loss in accuracy. Conclusion Our findings demonstrate that HiSAN-generated embeddings from SEER data can be used effectively in cancer site classification without significant demographic bias influencing the outcomes. Post-training pruning therefore functions as a practical audit and sensitivity check.

Shivanna, Abhishek [ORNL] (ORCID:0009000665228593)↗
DOE OSTI2026

Data for Promoter Deletion in the Soybean Compact Mutant Leads to Overexpression of a Gene with Homology to the C20-Gibberellin 2-Oxidase Family

Height is a critical component of plant architecture, significantly affecting crop yield. The genetic basis of this trait in soybean remains unclear. In this study, we report the characterization of the Compact mutant of soybean, which has short internodes. The candidate gene was mapped to chromosome 17, and the interval containing the causative mutation was further delineated using biparental mapping. Whole-genome sequencing of the mutant revealed an 8.7 kb deletion in the promoter of the Glyma.17g145200 gene, which encodes a member of the class III gibberellin (GA) 2-oxidases. The mutation has a dominant effect, likely via increased expression of the GA 2-oxidase transcript observed in green tissue, as a result of the deletion in the promoter of Glyma.17g145200. We further demonstrate that levels of GA precursors are altered in the Compact mutant, supporting a role in GA metabolism, and that the mutant phenotype can be rescued with exogenous GA3. We also determined that overexpression of Glyma.17g145200 in Arabidopsis results in dwarfed plants. Thus, gain of promoter activity in the Compact mutant leads to a short internode phenotype in soybean through altered metabolism of gibberellin precursors. These results provide an example of how structural variation can control an important crop trait and a role for Glyma.17g145200 in soybean architecture, with potential implications for increasing crop yield.

Biomass Analytics↗
DOE OSTI2026

Data for Expression of a Bacterial Trehalose 6-Phosphate Synthase Gene otsA in Camelina sativa Seeds Promotes the Channelling of Carbon Towards Oil Accumulation

Improving seed oil yield is essential for developing Camelina sativa as a sustainable biofuel crop. Fatty acid synthesis depends on the production of acetyl-CoA from photosynthetically derived sugars. Trehalose 6-phosphate (T6P), a proxy for sucrose availability, can link sugar status to plant growth and development. Synthesised by trehalose 6-phosphate synthase (TPS) from UDP-glucose and glucose-6-phosphate, T6P plays a regulatory role in metabolism. Our previous studies on Arabidopsis transgenic lines constitutively expressing the E. coli otsA (encoding TPS) showed increased T6P levels and seed triacylglycerol, along with stunted growth. In the present study we express otsA in camelina under the control of a seed-specific Phaseolin promoter. Seeds of the resulting transgenic lines accumulated high levels of T6P, and a 15%–20% increase in total fatty acids and triacylglycerol compared to wild-type. Molecular analysis showed the transgenic seeds had reduced SnRK1 activity, elevated WRI1 protein levels, and increased the levels of WRI1 and its target genes, along with enhanced rates of fatty acid synthesis that increased seed weights relative to wild type. Notably, the increase in oil did not affect seed protein levels but did reduce the soluble metabolite fraction. Crucially, seed-specific expression of otsA mitigated the growth defects associated with constitutive otsA expression, and the transgenic lines showed normal seed development and germination. These findings demonstrate that targeted T6P modulation via seed-specific otsA expression is an effective metabolic engineering strategy to boost oil production in camelina and potentially in other oilseed crops and bioenergy crops such as energycane, sorghum and miscanthus.

Lipids↗
DOE OSTI2026

Data for Sugar Accumulation Enhancement in Sorghum Stem is Associated with Reduced Reproductive Sink Strength and Increased Phloem Unloading Activity

Sweet sorghum has emerged as a promising source of bioenergy mainly due to its high biomass and high soluble sugar yield in stems. Studies have shown that loss-of-function Dry locus alleles have been selected during sweet sorghum domestication, and decapitation can further boost sugar accumulation in sweet sorghum, indicating that the potential for improving sugar yields is yet to be fully realized. To maximize sugar accumulation, it is essential to gain a better understanding of the mechanism underlying the massive accumulation of soluble sugars in sweet sorghum stems in addition to the Dry locus. We performed a transcriptomic analysis upon decapitation of near-isogenic lines for mutant (d, juicy stems, and green leaf midrib) and functional (D, dry stems and white leaf midrib) alleles at the Dry locus. Our analysis revealed that decapitation suppressed photosynthesis in leaves, but accelerated starch metabolic processes in stems. SbbHLH093 negatively correlates with sugar levels supported by genotypes (DD vs. dd), treatments (control vs. decapitation), and developmental stages post anthesis (3d vs.10d). D locus gene SbNAC074A and other programmed cell death-related genes were down regulated by decapitation, while sugar transporter-encoding gene SbSWEET1A was induced. Both SbSWEET1A and Invertase 5 were detected in phloem companion cells by RNA in situ assay. Loss of the SbbHLH093 homolog, AtbHLH093, in Arabidopsis led to a sugar accumulation increase. This study provides new insights into sugar accumulation enhancement in bioenergy crops, which can be potentially achieved by reducing reproductive sink strength and enhancing phloem unloading.

Transcriptomics↗
DOE OSTI2026

Data for A Role for Differential Rubisco Activase Isoform Expression in C4 Bioenergy Grasses at High Temperature

Rubisco activase (Rca) facilitates the release of sugar-phosphate inhibitors at Rubisco catalytic sites during CO2 fixation. Most plant species express two Rca isoforms, the larger Rca-α and the shorter Rca-β, either by alternative splicing from a single gene or expression from separate genes. The mechanism of Rubisco activation by Rca isoforms has been intensively studied in C3 plants. However, the functional role of Rca in C4 plants where Rubisco and Rca are located in a much higher [CO2] compartment is less clear. In this study, we selected four C4 bioenergy grasses and the model C4 grass setaria ( Setaria viridis ) to investigate the role of Rca in C4 photosynthesis. All five C4 grass species contained two Rca genes, one encoding Rca-α and the other Rca-β, which were positioned closely together in the genomes. A variety of abiotic stress-related motifs were identified in the Rca-α promoter of each grass, and while the Rca-β gene was constantly highly expressed at ambient temperature, Rca-α isoforms were expressed only at high temperature but never surpassed 30% of Rca-β content. The pattern of Rca-α induction on transition to high temperature and reduction on return to ambient temperature was the same in all five C4 grasses. In sorghum ( Sorghum bicolor ), sugarcane ( Saccharum officinarum ), and setaria, the induction rate of Rca-α was similar to the recovery rate of photosynthesis and Rubisco activation at high temperature. This association between Rca-α isoform expression and maintenance of Rubisco activation at high temperature suggests that Rca-α has a functional thermo-protective role in carbon fixation in C4 grasses by sustaining Rubisco activation at high temperature.

Genomics↗
DOE OSTI2020

Exact design of complex amplitude holograms for producing arbitrary scalar fields

Typical methods to holographically encode arbitrary wavefronts assume the hologram medium only applies either phase shifts or amplitude attenuation to the wavefront. In many cases, phase cannot be introduced to the wavefront without also affecting the amplitude. Here we show how to encode an arbitrary wavefront into an off-axis transmission hologram that returns the exact desired arbitrary wavefunction in a diffracted beam for phase-only, amplitude-only, or mixed phase and amplitude holograms with any periodic groove profile. We apply this to design thin holograms for electrons in a TEM, but our results are generally applicable to light and X-ray optics. We employ a phase reconstruction from a series of focal plane images to qualitatively show the accuracy of this method to impart the expected amplitude and phase to a specific diffraction order.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2022

Generation of polarization entanglement via the quantum Zeno effect

The quantum Zeno effect reveals that continuous observation of a quantum system can significantly alter its evolution. Here, we present a method for establishing polarization entanglement between two initially unentangled photons in coupled waveguides via the quantum Zeno effect. We support our analytical investigation with numerical simulations of the underlying Schrodinger equation describing the system. Further, we extend our technique to three coupled waveguides in a planar configuration and determine the parameters required to generate three-qubit W-states. In contrast to existing schemes based on a vacuum and single-photon encoding, the polarization encoding in our approach is compatible with quantum information protocols that remove photon loss through post-selection. Our findings offer a powerful quantum state engineering approach for photonic quantum information technologies.

Nodurft, Ian C.↗
DOE OSTI2024

Reconstructing Richtmyer–Meshkov instabilities from noisy radiographs using low dimensional features and attention-based neural networks

We develop an ML-based approach for density reconstruction based on transformer neural networks. This approach is demonstrated in the setting of ICF-like double shell hydrodynamic simulations wherein the parameters related to material properties and initial conditions are varied. The new method can robustly recover the complex topologies given by the Richtmyer-Meshkoff instability (RMI) from a sequence of hydrodynamic features derived from radiographic images corrupted with blur, scatter, and noise. A noise model is developed to characterize errors in extracting features from synthetic radiographs of the simulated density field. The key component of the network is a transformer encoder that acts on a sequence of features extracted from noisy radiographs. This encoder includes numerous self-attention layers that act to learn temporal dependencies in the input sequences and increase the expressiveness of the model. This approach is shown to exhibit an excellent ability to accurately recover the RMI growth rates, despite the gas-metal interface being greatly obscured by radiographic noise. Our approach can be applied in a broad array of fields involving shock physics and material science.

47 OTHER INSTRUMENTATION↗
DOE OSTI2023

Frequency-bin photonic quantum information

Discrete frequency modes, or bins, present a blend of opportunities and challenges for photonic quantum information processing. Frequency-bin-encoded photons are readily generated by integrated quantum light sources, naturally high-dimensional, stable in optical fiber, and massively parallelizable in a single spatial mode. Yet quantum operations on frequency-bin states require coherent and controllable multifrequency interference, making them significantly more challenging to manipulate than more traditional spatial degrees of freedom. In this mini-review, we describe recent developments that have transformed these challenges and propelled frequency bins forward. Focusing on sources, manipulation schemes, and detection approaches, we introduce the basics of frequency-bin encoding, summarize the state of the art, and speculate on the field’s next phases. Given the combined progress in integrated photonics, high-fidelity quantum gates, and proof-of-principle demonstrations, frequency-bin quantum information is poised to emerge from the lab and leave its mark on practical quantum information processing—particularly in networking where frequency bins offer unique tools for multiplexing, interconnects, and high-dimensional communications.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗
DOE OSTI2024

High-dimensional quantum key distribution using orbital angular momentum of single photons from a colloidal quantum dot at room temperature

High-dimensional quantum key distribution (HDQKD) is a promising avenue to address the inherent limitations of basic quantum key distribution (QKD) protocols. However, experimental realizations of HDQKD to date have relied on indeterministic photon sources that limit the achievable key rate. In this paper, we demonstrate a full emulation of a HDQKD system using a single colloidal giant quantum dot (gQD) as a deterministic, compact, and room-temperature single-photon source (SPS). We demonstrate a practical protocol by encoding information in a high-dimensional space (d = 3) of the orbital angular momentum of the photons. Our experimental configuration incorporates two spatial light modulators for encoding and decoding the spatial information carried by individual photons. Our experimental demonstration establishes the feasibility of utilizing high radiative quantum yield gQDs as practical SPSs for HDQKD. We also experimentally demonstrate surpassing the traditional d = 2 QKD capacity with comparable error rates, indicating a significant improvement in performance while maintaining reliability.

36 MATERIALS SCIENCE↗
DOE OSTI2021

The novel anti-CRISPR AcrIIA22 relieves DNA torsion in target plasmids and impairs SpyCas9 activity

To overcome CRISPR-Cas defense systems, many phages and mobile genetic elements (MGEs) encode CRISPR-Cas inhibitors called anti-CRISPRs (Acrs). Nearly all characterized Acrs directly bind Cas proteins to inactivate CRISPR immunity. Here, using functional metagenomic selection, we describe AcrIIA22, an unconventional Acr found in hypervariable genomic regions of clostridial bacteria and their prophages from human gut microbiomes. AcrIIA22 does not bind strongly to SpyCas9 but nonetheless potently inhibits its activity against plasmids. To gain insight into its mechanism, we obtained an X-ray crystal structure of AcrIIA22, which revealed homology to PC4-like nucleic acid–binding proteins. Based on mutational analyses and functional assays, we deduced that acrIIA22 encodes a DNA nickase that relieves torsional stress in supercoiled plasmids. This may render them less susceptible to SpyCas9, which uses free energy from negative supercoils to form stable R-loops. Modifying DNA topology may provide an additional route to CRISPR-Cas resistance in phages and MGEs.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2025

The role of chromatin state in intron retention: A case study in leveraging large scale deep learning models

Complex deep learning models trained on very large datasets have become key enabling tools for current research in natural language processing and computer vision. By providing pre-trained models that can be fine-tuned for specific applications, they enable researchers to create accurate models with minimal effort and computational resources. Large scale genomics deep learning models come in two flavors: the first are large language models of DNA sequences trained in a self-supervised fashion, similar to the corresponding natural language models; the second are supervised learning models that leverage large scale genomics datasets from ENCODE and other sources. We argue that these models are the equivalent of foundation models in natural language processing in their utility, as they encode within them chromatin state in its different aspects, providing useful representations that allow quick deployment of accurate models of gene regulation. We demonstrate this premise by leveraging the recently created Sei model to develop simple, interpretable models of intron retention, and demonstrate their advantage over models based on the DNA language model DNABERT-2. Our work also demonstrates the impact of chromatin state on the regulation of intron retention. Using representations learned by Sei, our model is able to discover the involvement of transcription factors and chromatin marks in regulating intron retention, providing better accuracy than a recently published custom model developed for this purpose.

Biochemistry & Molecular Biology↗
DOE OSTI2024

Contact-dependent growth inhibition (CDI) systems deploy a large family of polymorphic ionophoric toxins for inter-bacterial competition

Contact-dependent growth inhibition (CDI) is a widespread form of inter-bacterial competition mediated by CdiA effector proteins. CdiA is presented on the inhibitor cell surface and delivers its toxic C-terminal region (CdiA-CT) into neighboring bacteria upon contact. Inhibitor cells also produce CdiI immunity proteins, which neutralize CdiA-CT toxins to prevent auto-inhibition. Here, we describe a diverse group of CDI ionophore toxins that dissipate the transmembrane potential in target bacteria. These CdiA-CT toxins are composed of two distinct domains based on AlphaFold2 modeling. The C-terminal ionophore domains are all predicted to form five-helix bundles capable of spanning the cell membrane. The N-terminal "entry" domains are variable in structure and appear to hijack different integral membrane proteins to promote toxin assembly into the lipid bilayer. The CDI ionophores deployed by E. coli isolates partition into six major groups based on their entry domain structures. Comparative sequence analyses led to the identification of receptor proteins for ionophore toxins from groups 1 & 3 (AcrB), group 2 (SecY) and groups 4 (YciB). Using forward genetic approaches, we identify novel receptors for the group 5 and 6 ionophores. Group 5 exploits homologous putrescine import proteins encoded by puuP and plaP, and group 6 toxins recognize di/tripeptide transporters encoded by paralogous dtpA and dtpB genes. Finally, we find that the ionophore domains exhibit significant intra-group sequence variation, particularly at positions that are predicted to interact with CdiI. Accordingly, the corresponding immunity proteins are also highly polymorphic, typically sharing only ~30% sequence identity with members of the same group. Competition experiments confirm that the immunity proteins are specific for their cognate ionophores and provide no protection against other toxins from the same group. The specificity of this protein interaction network provides a mechanism for self/nonself discrimination between E. coli isolates.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2020

Distinctiveness of genes contributing to growth of Pseudomonas syringae in diverse host plant species

A variety of traits are necessary for bacterial colonization of the interior of plant hosts, including well-studied virulence effectors as well as other phenotypes contributing to bacterial growth and survival within the apoplast. High-throughput methods such as transposon sequencing (TnSeq) are powerful tools to identify such genes in bacterial pathogens. However, there is little information as to the distinctiveness of traits required for bacterial colonization of different hosts. Here, we utilize randomly barcoded TnSeq (RB-TnSeq) to identify the genes that contribute to the ability of Pseudomonas syringae strain B728a to grow within common bean (Phaseolus vulgaris), lima bean (Phaseolus lunatus), and pepper (Capsicum annuum); species representing two different plant families. The magnitude of contribution of most genes to apoplastic fitness in each of the plant hosts was similar. However, 50 genes significantly differed in their fitness contributions to growth within these species. These genes encoded proteins in various functional categories including polysaccharide synthesis and transport, amino acid metabolism and transport, cofactor metabolism, and phytotoxin synthesis and transport. Six genes that encoded unannotated, hypothetical proteins also contributed differentially to growth in these hosts. The genetic repertoire of a relatively promiscuous pathogen such as P. syringae may thus be shaped, at least in part, by the conditional contribution of some fitness determinants.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2022

Classification of bacterial plasmid and chromosome derived sequences using machine learning

Plasmids are important genetic elements that facilitate horizonal gene transfer between bacteria and contribute to the spread of virulence and antimicrobial resistance. Most bacterial genome sequences in the public archives exist in draft form with many contigs, making it difficult to determine if a contig is of chromosomal or plasmid origin. Using a training set of contigs comprising 10,584 chromosomes and 10,654 plasmids from the PATRIC database, we evaluated several machine learning models including random forest, logistic regression, XGBoost, and a neural network for their ability to classify chromosomal and plasmid sequences using nucleotide k-mers as features. Based on the methods tested, a neural network model that used nucleotide 6-mers as features that was trained on randomly selected chromosomal and plasmid subsequences 5kb in length achieved the best performance, outperforming existing out-of-the-box methods, with an average accuracy of 89.38% ± 2.16% over a 10-fold cross validation. The model accuracy can be improved to 92.08% by using a voting strategy when classifying holdout sequences. In both plasmids and chromosomes, subsequences encoding functions involved in horizontal gene transfer—including hypothetical proteins, transporters, phage, mobile elements, and CRISPR elements—were most likely to be misclassified by the model. This study provides a straightforward approach for identifying plasmid-encoding sequences in short read assemblies without the need for sequence alignment-based tools.

59 BASIC BIOLOGICAL SCIENCES↗