Search NASA⌕ Search

SEARCH · Search NASA

Results for “sequencing”

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 55 records · Page 3
All sourcesNASA NTRSDOE OSTI
NASA NTRS2005

Rover Sequencing and Visualization Program

The Rover Sequencing and Visualization Program (RSVP) is the software tool for use in the Mars Exploration Rover (MER) mission for planning rover operations and generating command sequences for accomplishing those operations. RSVP combines three-dimensional (3D) visualization for immersive exploration of the operations area, stereoscopic image display for high-resolution examination of the downlinked imagery, and a sophisticated command-sequence editing tool for analysis and completion of the sequences. RSVP is linked with actual flight-code modules for operations rehearsal to provide feedback on the expected behavior of the rover prior to committing to a particular sequence. Playback tools allow for review of both rehearsed rover behavior and downlinked results of actual rover operations. These can be displayed simultaneously for comparison of rehearsed and actual activities for verification. The primary inputs to RSVP are downlink data products from the Operations Storage Server (OSS) and activity plans generated by the science team. The activity plans are high-level goals for the next day s activities. The downlink data products include imagery, terrain models, and telemetered engineering data on rover activities and state. The Rover Sequence Editor (RoSE) component of RSVP performs activity expansion to command sequences, command creation and editing with setting of command parameters, and viewing and management of rover resources. The HyperDrive component of RSVP performs 2D and 3D visualization of the rover s environment, graphical and animated review of rover-predicted and telemetered state, and creation and editing of command sequences related to mobility and Instrument Deployment Device (IDD) operations. Additionally, RoSE and HyperDrive together evaluate command sequences for potential violations of flight and safety rules. The products of RSVP include command sequences for uplink that are stored in the Distributed Object Manager (DOM) and predicted rover state histories stored in the OSS for comparison and validation of downlinked telemetry. The majority of components comprising RSVP utilize the MER command and activity dictionaries to automatically customize the system for MER activities. Thus, RSVP, being highly data driven, may be tailored to other missions with minimal effort. In addition, RSVP uses a distributed, message-passing architecture to allow multitasking, and collaborative visualization and sequence development by scattered team members.

Cooper, Brian↗
NASA NTRS2017

Exploring Connectivity in Sequence Space of Functional RNA

Emergence of replicable genetic molecules was one of the marking points in the origin of life, evolution of which can be conceptualized as a walk through the space of all possible sequences. A theoretical concept of fitness landscape helps to understand evolutionary processes through assigning a value of fitness to each genotype. Then, evolution of a phenotype is viewed as a series of consecutive, single-point mutations. Natural selection biases evolution toward peaks of high fitness and away from valleys of low fitness. whereas neutral drift occurs in the sequence space without direction as mutations are introduced at random. Large networks of neutral or near-neutral mutations on a fitness landscape, especially for sufficiently long genomes, are possible or even inevitable. Their detection in experiments, however, has been elusive. Although a few near-neutral evolutionary pathways have been found, recent experimental evidence indicates landscapes consist of largely isolated islands. The generality of these results, however, is not clear, as the genome length or the fraction of functional molecules in the genotypic space might have been insufficient for the emergence of large, neutral networks. Thorough investigation on the structure of the fitness landscape is essential to understand the mechanisms of evolution of early genomes. RNA molecules are commonly assumed to play the pivotal role in the origin of genetic systems. They are widely believed to be early, if not the earliest, genetic and catalytic molecules, with abundant biochemical activities as aptamers and ribozymes, i.e. RNA molecules capable, respectively, to bind small molecules or catalyze chemical reactions. Here, we present results of our recent studies on the structure of the sequence space of RNA ligase ribozymes selected through in vitro evolution. Several hundred thousands of sequences active to a different degree were obtained by way of deep sequencing. Analysis of these sequences revealed several large clusters defined such that every sequence in a cluster can be reached from any other sequence in the same cluster through a series of single point mutations. Sequences in a single cluster appear to adopt more than one secondary structure. The mechanism of refolding within a single cluster was examined. To shed light on possible evolutionary paths in the space of ribozymes, the connectivity between clusters was investigated. The effect of length of RNA molecules on the structure of the fitness landscape and possible evolutionary paths was examined by way of comparing functional sequences of 20 and 80 nucleobases in length. It was found that sequences of different lengths shared secondary structure motifs that were presumed responsible for catalytic activity, with increasing complexity and global structural rearrangements emerging in longer molecules.

Wei, Chenyu↗
NASA NTRS2021

Reducing Risk of InSight Surface Operations Through High-Fidelity Command Sequence Modeling

Simulating spacecraft behavior is crucial for the success of deep space missions, and failure to do so may result in damages to or the loss of the spacecraft. Many previous deep space missions have made use of ground-simulation of sequenced commanding, at speeds far greater than real time, to predict spacecraft state over time through the execution of onboard sequences. This type of modeling can be done at any fidelity, and most missions have opted to decrease fidelity to reduce cost and complexity. However, NASA’s Interior Exploration using Seismic Investigations, Geodesy and Heat Transport (InSight) mission expanded the scope of ground modeling considerably, which has led to numerous benefits over past implementations. This paper will discuss the process and products that InSight created, as well as the lessons learned from successfully operating the spacecraft on Mars. InSight is the first JPL mission to expand the scope of ground modeling to include the uplink of files from Earth to the spacecraft, rather than making the simplification that any command sequences already exist onboard the spacecraft. The advantages of modeling the uplink of files are numerous. First, it allows for accurate modeling of the onboard filesystem of the spacecraft at all points in time, meaning that all file loads and deletions throughout the mission are modeled at the exact moment they are predicted to actually happen. Second, operators can be more certain that dependencies between sequences are not broken due to the dynamic nature of the filesystem as files are deleted, copied, and uplinked. Lastly, spacecraft filesystem tracking allows for management of sequences prior to uplink, limiting the uplink to only new sequences. The onboard filesystem model became crucial to mission success, emphasizing the importance of investing in accurate models before the need for them arises. During daily tactical operations of a spacecraft on Mars, a model is only useful if the results can be interpreted quickly. In this fast-paced environment, it is essential that command products are modeled and reviewed, errors are found and diagnosed, and new command products are redelivered, remodeled, re-reviewed in a timely manner. It is impossible to review the entire model and therefore the results of the model must be condensed and presented in a fashion that is intuitive, easy-to-navigate, complete, and trustworthy. InSight developed a number of innovative sequence review products that are designed to provide operators with the information required to quickly assess the validity of command products and diagnose potential issues. Together, these products provide a complete, yet succinct picture of the command and sequence model to the operators and facilitate a quick assessment of all sequence command products. This paper will cover planning and sequencing innovations made during InSight surface operations, and will compare the tools, processes, and results to those on other missions. Additionally, the paper will cover the flexible, yet robust nature of the planning and sequencing system architecture and how that flexibility allowed for rapid development and response to the unpredictability of Mars.

Cloutier, Kyle↗
NASA NTRS1998

Analysis of xylem formation in pine by cDNA sequencing

Secondary xylem (wood) formation is likely to involve some genes expressed rarely or not at all in herbaceous plants. Moreover, environmental and developmental stimuli influence secondary xylem differentiation, producing morphological and chemical changes in wood. To increase our understanding of xylem formation, and to provide material for comparative analysis of gymnosperm and angiosperm sequences, ESTs were obtained from immature xylem of loblolly pine (Pinus taeda L.). A total of 1,097 single-pass sequences were obtained from 5' ends of cDNAs made from gravistimulated tissue from bent trees. Cluster analysis detected 107 groups of similar sequences, ranging in size from 2 to 20 sequences. A total of 361 sequences fell into these groups, whereas 736 sequences were unique. About 55% of the pine EST sequences show similarity to previously described sequences in public databases. About 10% of the recognized genes encode factors involved in cell wall formation. Sequences similar to cell wall proteins, most known lignin biosynthetic enzymes, and several enzymes of carbohydrate metabolism were found. A number of putative regulatory proteins also are represented. Expression patterns of several of these genes were studied in various tissues and organs of pine. Sequencing novel genes expressed during xylem formation will provide a powerful means of identifying mechanisms controlling this important differentiation pathway.

Non-NASA Center↗
DOE OSTI2026

Circularity in Sequence-Controlled Copolyamides Enabled by Regioselective Enzymatic Hydrolysis

Sequence-controlled polymers enable precise control over macromolecular structures and function, but both their synthesis and end-of-life management remain fundamental challenges. Achieving high sequence fidelity is synthetically demanding, and conventional depolymerization methods lack regioselectivity, leading to irreversible loss of encoded molecular information and limiting polymer circularity. Enzymatic catalysis offers a potential solution by combining substrate specificity with selective bond cleavage. Here, we report the synthesis, characterization, and regioselective enzymatic depolymerization of poly- (X,AMA), a sequence-controlled copolyamide composed of alternating hexamethylenediamine−adipic acid (MA) and pxylylenediamine− adipic acid (XA) repeat units. Poly(X,AMA) was synthesized via solid-state polycondensation (SSP) of sequence-defined oligomers, enabling precise control over repeat-unit order. Polymer microstructure and sequence fidelity were confirmed by 13 C NMR spectroscopy and MALDI−TOF mass spectrometry. Comparison with a statistical copolymer analogue and Nylon-66 demonstrated pronounced differences in crystallinity, morphology, and thermal behavior arising from sequence control. Screening of 96 Nylon hydrolase homologues against poly(X,AMA) revealed strongly enzyme-dependent depolymerization profiles. While tetrad formation was generally favored, enzymes displayed pronounced sequence selectivity, preferentially releasing distinct sequence-defined tetrads XAMA or MAXA. SSP of sequence-defined tetrad MAXA produced a copolyamide with near identical monomer ordering as poly(X,AMA). Computational modeling of enzyme−substrate complexes identified structural features consistent with the observed regioselectivity. Together, these results establish selective enzymatic depolymerization as a viable strategy for the circular recycling of sequence-controlled polymers and provide a foundation for the rational engineering of enzymes for programmable polymer deconstruction.

Amides↗
NASA NTRS2002

Identification of characteristic oligonucleotides in the bacterial 16S ribosomal RNA sequence dataset

MOTIVATION: The phylogenetic structure of the bacterial world has been intensively studied by comparing sequences of 16S ribosomal RNA (16S rRNA). This database of sequences is now widely used to design probes for the detection of specific bacteria or groups of bacteria one at a time. The success of such methods reflects the fact that there are local sequence segments that are highly characteristic of particular organisms or groups of organisms. It is not clear, however, the extent to which such signature sequences exist in the 16S rRNA dataset. A better understanding of the numbers and distribution of highly informative oligonucleotide sequences may facilitate the design of hybridization arrays that can characterize the phylogenetic position of an unknown organism or serve as the basis for the development of novel approaches for use in bacterial identification. RESULTS: A computer-based algorithm that characterizes the extent to which any individual oligonucleotide sequence in 16S rRNA is characteristic of any particular bacterial grouping was developed. A measure of signature quality, Q(s), was formulated and subsequently calculated for every individual oligonucleotide sequence in the size range of 5-11 nucleotides and for 15mers with reference to each cluster and subcluster in a 929 organism representative phylogenetic tree. Subsequently, the perfect signature sequences were compared to the full set of 7322 sequences to see how common false positives were. The work completed here establishes beyond any doubt that highly characteristic oligonucleotides exist in the bacterial 16S rRNA sequence dataset in large numbers. Over 16,000 15mers were identified that might be useful as signatures. Signature oligonucleotides are available for over 80% of the nodes in the representative tree.

NASA Discipline Life Sciences Technologies↗
NASA NTRS2008

MRO Sequence Checking Tool

The MRO Sequence Checking Tool program, mro_check, automates significant portions of the MRO (Mars Reconnaissance Orbiter) sequence checking procedure. Though MRO has similar checks to the ODY s (Mars Odyssey) Mega Check tool, the checks needed for MRO are unique to the MRO spacecraft. The MRO sequence checking tool automates the majority of the sequence validation procedure and check lists that are used to validate the sequences generated by MRO MPST (mission planning and sequencing team). The tool performs more than 50 different checks on the sequence. The automation varies from summarizing data about the sequence needed for visual verification of the sequence, to performing automated checks on the sequence and providing a report for each step. To allow for the addition of new checks as needed, this tool is built in a modular fashion.

Fisher, Forest↗
NASA NTRS2017

Novel Approach to Quantification of Telomere Length with Direct Nanopore Sequencing and PCR Amplification

The ends of human chromosomes contain telomeres, or tandem arrays of repeating DNA sequences capped by multiple associated proteins that protect chromosomal ends from degradation. Telomeres function to preserve genomic stability by preventing natural chromosomal ends from being recognized as broken DNA double-strand breaks and triggering inappropriate DNA damage responses. Mounting evidence shows telomere length is an inherited trait that decreases with cellular division and normal aging. In addition, telomere length also appears to be influenced by other factors such as cellular oxidative stress, radiation and mechanical unloading of tissues as in microgravity. To measure these potential effects of the space environment on telomere lengths and cellular aging and regenerative potential we developed a novel telomere measurement approach based on nanopore sequencing of PCR amplified bar-coded chromosome termini. Specifically, telomeres can be directly enriched using barcode sequences ligated to the end of a free end- repaired telomere using the WetLab-2 facility SmartCycler on ISS. Prior to the ligation and amplification protocol a proteinase K digestion of capping proteins followed by a single 95-degree C heat denaturation of the protease is included. After digestion and bar-code ligation, PCR amplification will initiate with the ligated barcoded sequence, suppressing amplification of intra-genomic fragments and resulting in long read barcoded telomere amplicons including the nanopore motor protein sequences. Purified PCR amplicons are then used for nanopore sequencing library generation by simple addition of motor proteins and sequencing library is loaded into the MinION nanopore DNA-sequencer. Amplicon sequence reads from the nanopore device can be base-called quickly on ISS due to barcoding ligation and subsequent PCR amplification enhancing the telomere sequence resolution. If successfully implemented on ISS this technique will provide a novel means of measuring regenerative ability of somatic stem cells in astronauts, and of determining whether spaceflight in microgravity alters their telomere lengths and causes premature cellular aging.

Ma, Kristin R.↗
NASA NTRS1995

On the joint spectral density of bivariate random sequences

For univariate random sequences, the power spectral density acts like a probability density function of the frequencies present in the sequence. This dissertation extends that concept to bivariate random sequences. For this purpose, a function called the joint spectral density is defined that represents a joint probability weighing of the frequency content of pairs of random sequences. Given a pair of random sequences, the joint spectral density is not uniquely determined in the absence of any constraints. Two approaches to constraining the sequences are suggested: (1) assume the sequences are the margins of some stationary random field, (2) assume the sequences conform to a particular model that is linked to the joint spectral density. For both approaches, the properties of the resulting sequences are investigated in some detail, and simulation is used to corroborate theoretical results. It is concluded that under either of these two constraints, the joint spectral density can be computed from the non-stationary cross-correlation.

Aalfs, David D.↗
NASA NTRS2013

Robot Sequencing and Visualization Program (RSVP)

The Robot Sequencing and Visualization Program (RSVP) is being used in the Mars Science Laboratory (MSL) mission for downlink data visualization and command sequence generation. RSVP reads and writes downlink data products from the operations data server (ODS) and writes uplink data products to the ODS. The primary users of RSVP are members of the Rover Planner team (part of the Integrated Planning and Execution Team (IPE)), who use it to perform traversability/articulation analyses, take activity plan input from the Science and Mission Planning teams, and create a set of rover sequences to be sent to the rover every sol. The primary inputs to RSVP are downlink data products and activity plans in the ODS database. The primary outputs are command sequences to be placed in the ODS for further processing prior to uplink to each rover. RSVP is composed of two main subsystems. The first, called the Robot Sequence Editor (RoSE), understands the MSL activity and command dictionaries and takes care of converting incoming activity level inputs into command sequences. The Rover Planners use the RoSE component of RSVP to put together command sequences and to view and manage command level resources like time, power, temperature, etc. (via a transparent realtime connection to SEQGEN). The second component of RSVP is called HyperDrive, a set of high-fidelity computer graphics displays of the Martian surface in 3D and in stereo. The Rover Planners can explore the environment around the rover, create commands related to motion of all kinds, and see the simulated result of those commands via its underlying tight coupling with flight navigation, motor, and arm software. This software is the evolutionary replacement for the Rover Sequencing and Visualization software used to create command sequences (and visualize the Martian surface) for the Mars Exploration Rover mission.

Cooper, Brian K.↗
NASA NTRS2011

Project Report: Automatic Sequence Processor Software Analysis

The Mission Planning and Sequencing (MPS) element of Multi-Mission Ground System and Services (MGSS) provides space missions with multi-purpose software to plan spacecraft activities, sequence spacecraft commands, and then integrate these products and execute them on spacecraft. Jet Propulsion Laboratory (JPL) is currently is flying many missions. The processes for building, integrating, and testing the multi-mission uplink software need to be improved to meet the needs of the missions and the operations teams that command the spacecraft. The Multi-Mission Sequencing Team is responsible for collecting and processing the observations, experiments and engineering activities that are to be performed on a selected spacecraft. The collection of these activities is called a sequence and ultimately a sequence becomes a sequence of spacecraft commands. The operations teams check the sequence to make sure that no constraints are violated. The workflow process involves sending a program start command, which activates the Automatic Sequence Processor (ASP). The ASP is currently a file-based system that is comprised of scripts written in perl, c-shell and awk. Once this start process is complete, the system checks for errors and aborts if there are any; otherwise the system converts the commands to binary, and then sends the resultant information to be radiated to the spacecraft.

sequencing↗
NASA NTRS1994

Transcription factor IID in the Archaea: sequences in the Thermococcus celer genome would encode a product closely related to the TATA-binding protein of eukaryotes

The first step in transcription initiation in eukaryotes is mediated by the TATA-binding protein, a subunit of the transcription factor IID complex. We have cloned and sequenced the gene for a presumptive homolog of this eukaryotic protein from Thermococcus celer, a member of the Archaea (formerly archaebacteria). The protein encoded by the archaeal gene is a tandem repeat of a conserved domain, corresponding to the repeated domain in its eukaryotic counterparts. Molecular phylogenetic analyses of the two halves of the repeat are consistent with the duplication occurring before the divergence of the archael and eukaryotic domains. In conjunction with previous observations of similarity in RNA polymerase subunit composition and sequences and the finding of a transcription factor IIB-like sequence in Pyrococcus woesei (a relative of T. celer) it appears that major features of the eukaryotic transcription apparatus were well-established before the origin of eukaryotic cellular organization. The divergence between the two halves of the archael protein is less than that between the halves of the individual eukaryotic sequences, indicating that the average rate of sequence change in the archael protein has been less than in its eukaryotic counterparts. To the extent that this lower rate applies to the genome as a whole, a clearer picture of the early genes (and gene families) that gave rise to present-day genomes is more apt to emerge from the study of sequences from the Archaea than from the corresponding sequences from eukaryotes.

NASA Discipline Exobiology↗
DOE OSTI2024

Development of near-optimal advanced control sequences for chiller plants with water-side economizers in U.S. Climates (ASHRAE RP-1661)

Various advanced control sequences for chiller plants with water-side economizers (WSE) have been proposed in literature, but the evaluation and optimization of those controls is limited. It is possible to maximize energy savings by selecting different sequences and related parameters based on the plant configuration, load, and climate. This paper addresses this gap by developing near-optimal advanced control sequences for chiller plants with WSEs. First, advanced control sequences for chiller plants with WSEs are categorized into condenser water, chilled water, and hybrid controls and representative sequences from each category are identified. Next, 504 different scenarios are optimized. These scenarios represent all possible combinations of two plant configurations, a constant or variable load profile, three advanced control sequences, and seven optimization parameter combinations in six climate zones. The results show the recommended near-optimal sequences can reduce energy consumption by up to 15% relative to the baseline depending on the configuration, load profile, and climate. Specifically, the CW-CHW sequence is recommended for the majority of systems because it is often the most energy efficient and/or reduces the runtime of chillers. The methodology in this paper provides practical guidance for achieving energy savings through near-optimal control of chiller plants with WSEs.

42 ENGINEERING↗
NASA NTRS1986

A computer aided thermodynamic approach for predicting the formation of Z-DNA in naturally occurring sequences

The ease with which a particular DNA segment adopts the left-handed Z-conformation depends largely on the sequence and on the degree of negative supercoiling to which it is subjected. We describe a computer program (Z-hunt) that is designed to search long sequences of naturally occurring DNA and retrieve those nucleotide combinations of up to 24 bp in length which show a strong propensity for Z-DNA formation. Incorporated into Z-hunt is a statistical mechanical model based on empirically determined energetic parameters for the B to Z transition accumulated to date. The Z-forming potential of a sequence is assessed by ranking its behavior as a function of negative superhelicity relative to the behavior of similar sized randomly generated nucleotide sequences assembled from over 80,000 combinations. The program makes it possible to compare directly the Z-forming potential of sequences with different base compositions and different sequence lengths. Using Z-hunt, we have analyzed the DNA sequences of the bacteriophage phi X174, plasmid pBR322, the animal virus SV40 and the replicative form of the eukaryotic adenovirus-2. The results are compared with those previously obtained by others from experiments designed to locate Z-DNA forming regions in these sequences using probes which show specificity for the left-handed DNA conformation.

Non-NASA Center↗
DOE OSTI2025

Bottom-Up Simulation, Reconstruction, and Quantification of Macromolecule Sequences from Experimental Polymerizations

Motivated by the canonical sequence–structure–function paradigm, tools to characterize chemical patterning in natural biomacromolecules, from proteins to nucleic acids, have grown exponentially in recent years. However, analogous strategies for synthetic macromolecules remain in nascent stages, complicated by sequence polydispersity and analytical limitations. To address this, we have developed a comprehensive and open-source Python package, PRISM (polymer rate insights and sequence modeling), an end-to-end workflow that provides a path from experimental kinetics measurements to quantitative and qualitative metrics for describing chemical patterning in stochastic polymers. First, a numerical integration strategy was constructed to simulate and fit experimental data from reversible addition–fragmentation chain transfer (RAFT) polymerization kinetics, enabling the facile estimation of relevant reactivity ratios. These ratios were then used in a mechanism-specific stochastic kinetic simulation strategy to simulate sequence ensembles corresponding to model systems spanning experimental copolymers, classes of statistical polymers (e.g., alternating, block, and gradient), and multiblock copolymers. Lastly, inspired by sequence homology metrics from bioinformatics, we introduce visualization strategies and quantitative metrics to facilitate comparisons of different sequence ensembles. As the sequence–structure–function paradigm becomes increasingly central in de novo design of synthetic macromolecules, this toolkit provides a first step toward accurate and representative sequence description and featurization.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2024

Sequence-defined structural transitions by calcium-responsive proteins

Biopolymer sequences dictate their functions, and protein-based polymers are a promising platform to establish sequence–function relationships for novel biopolymers. To efficiently explore vast sequence spaces of natural proteins, sequence repetition is a common strategy to tune and amplify specific functions. This strategy is applied to repeats-in-toxin (RTX) proteins with calcium-responsive folding behavior, which stems from tandem repeats of the nonapeptide GGXGXDXUX in which X can be any amino acid and U is a hydrophobic amino acid. To determine the functional range of this nonapeptide, we modified a naturally occurring RTX protein that forms β-roll structures in the presence of calcium. Sequence modifications focused on calcium-binding turns within the repetitive region, including either global substitution of nonconserved residues or complete replacement with tandem repeats of a consensus nonapeptide GGAGXDTLY. Some sequence modifications disrupted the typical transition from intrinsically disordered random coils to folded β rolls, despite conservation of the underlying nonapeptide sequence. Proteins enriched with smaller, hydrophobic amino acids adopted secondary structures in the absence of calcium and underwent structural rearrangements in calcium-rich environments. In contrast, proteins with bulkier, hydrophilic amino acids maintained intrinsic disorder in the absence of calcium. In conclusion, these results indicate a significant role of nonconserved amino acids in calcium-responsive folding, thereby revealing a strategy to leverage sequences in the design of tunable, calcium-responsive biopolymers.

37 INORGANIC, ORGANIC, PHYSICAL, AND ANALYTICAL CH↗
DOE OSTI2025

From sequence to protein structure and conformational dynamics with artificial intelligence/machine learning

The 2024 Nobel Prize in Chemistry was awarded in part for de novo protein structure prediction using AlphaFold2, an artificial intelligence/machine learning (AI/ML) model trained on vast amounts of sequence and three-dimensional structure data. AlphaFold2 and related models, including RoseTTAFold and ESMFold, employ specialized neural network architectures driven by attention mechanisms to infer relationships between sequence and structure. At a fundamental level, these AI/ML models operate on the long-standing hypothesis that the structure of a protein is determined by its amino acid sequence. More recently, AlphaFold2 has been adapted for the prediction of multiple protein conformations by subsampling multiple sequence alignments. Herein, we provide an overview of the deterministic relationship between sequence and structure, which was hypothesized over half a century ago with profound implications for the biological sciences ever since. We postulate that protein conformational dynamics are also determined, at least in part, by amino acid sequence and that this relationship may be leveraged for construction of AI/ML models dedicated to predicting protein conformational ensembles. Accordingly, we describe a conceptual model architecture, which may be trained on sequence data in combination with conformationally sensitive structural information, coming primarily from nuclear magnetic resonance (NMR) spectroscopy. Notwithstanding certain limitations in this context, NMR offers abundant structural heterogeneity conducive to conformational ensemble prediction. As NMR and other data continue to accumulate, sequence-informed prediction of protein structural dynamics with AI/ML has the potential to emerge as a transformative capability across the biological sciences.

Artificial intelligence↗
DOE OSTI2025

Sequence length scaling in vision transformers for scientific images on frontier

Vision Transformers (ViTs) are pivotal for foundational models in scientific imagery, including Earth science applications, due to their capability to process large sequence lengths. While transformers for text have inspired scaling sequence lengths in ViTs, adapting these for ViTs introduces unique challenges. We develop distributed sequence parallelism for ViTs, enabling them to handle up to 1M tokens. Our approach, leveraging DeepSpeed-Ulysses and Long-Sequence-Segmentation with model sharding, is the first to apply sequence parallelism in ViT training, achieving a 94% batch scaling efficiency on 2,048 AMD-MI250X GPUs. Evaluating sequence parallelism in ViTs, particularly in models up to 10B parameters, highlighted substantial bottlenecks. We countered these with hybrid sequence, pipeline, and flash attention strategies, to scale beyond single GPU memory limits. Our method significantly enhances climate modeling accuracy by 20% in temperature predictions, marking the first training of a vision transformer model to convergence with a sequence length of 188K tokens, using full self-attention.

Tsaris, Aristeidis (aris) [ORNL] (ORCID:0000000277↗