Results for “Ground control” — Page 8

NASA NTRSDate not supplied

Designing Autonomy into Interfaces for Long-Duration Missions

As NASA develops technologies for long-duration crewed missions, we must understand how communication between ground control teams and astronauts differs from the current dynamic to adapt new concepts for long-duration mission operations. Today, ground control teams support astronauts with immediate availability to answer questions, resolve issues, and manage activities. In the near future, however, extended communication delays during long-duration missions will require astronauts to become more autonomous. As many of the responsibilities shift from the ground control teams to the astronauts on-board, the concept of operations must also change from how it functions today. With increased astronaut autonomy, software tools must be developed that enable efficient completion of mission tasks without increased mental workload. Designing software tools to facilitate crew autonomy requires development teams to know which data will enhance quick decision making while providing necessary context for situational awareness of systems being managed on-board. NASA's Autonomous Systems and Operations (ASO) team is presently developing a software interface tool, EXPRESS (EXpedite the PRocessing of Experiments for Space Station) 2.5 to enable a long-duration crew to schedule activities for and operate autonomous systems. This paper describes details of the integrated human factors approach that drove the design of the elements of the software tool, including self-managed scheduling, constraint-driven planning, autonomous system fault recovery, and recommended troubleshooting actions. Additionally, this paper will chronicle ASO modifications of the user interface after the team's first flight demonstration, how it was based on lessons learned during software development, as well as from crew feedback in order to develop the current version which will be demonstrated on ISS in 2021. In the upcoming ISS demonstration, the astronaut crew will be given scenarios for scheduling and operating autonomous system activities, including off-nominal scenarios and autonomous system recoveries. The demonstration of the EXPRESS 2.5 tool is a step towards improved levels of autonomy as our new journeys take us farther into space.

autonomy↗

NASA NTRS1994

Temperature prediction of space flight experiments by computer thermal analysis

Life sciences experiments are especially sensitive to temperature. A small temperature difference between otherwise identical samples can cause various differences in biological reaction rates. Knowledge of experimental temperatures and temperature histories help to distinguish the effects of microgravity and temperature on spaceflight experiments compared to ground based studies, and allow appropriate controls and sensitivity tests. Up to the present time, the Orbiter (Space Shuttle) has not generally provided temperature measurement instrumentation inside ambient lockers located in the Mid-deck of the Orbiter, or inside similar facilities such as Spacehab and Spacelab, but many pieces of hardware do have temperature recording capability. Most of these temperatures, however, have only been roughly measured or estimated. Such reported experimental temperatures, while accurate within a range of several degrees Celsius, are of limited utility to biological researchers. The temperature controlled lockers used in spaceflight, such as Commerical-Refrigeration Incubation Modules (C-R/IMs), severely reduce the mass and volume available for test samples and do not necessarily provide uniform thermal environments. While these test carriers avoid some of the experimental temperature variations of the ambient lockers, the number of samples which can be accommodated in these temperature controlled units is limited. In the present work, improved models of thermal prediction and control were sought. Temperatures are predicted by thermal analysis software using empirical temperatures recorded during STS-57. These temperatures are compared to data recorded throughout the mission using Ambient Temperature Recorders (ATRs) located within several payload lockers. Additional test cases are undertaken using controlled ground experiments to more precisely determine the reliability of the thermal model. The approach presented should increase the utility of various spaceflight carriers in the support of biological and material science research and ground control studies done in preparation for flight.

Birdsong, M. B.↗

NASA NTRSDate not supplied

Automated Software for Manned Spacecraft - Bridging the Gap from Sci Fi to Reality

With a voice command or a few taps on the console, the spacecraft pivots on a dime at high velocity and gently docks to an orbiting space platform. This is the image most people have of the complex software computations and integrated hardware performance necessary for a spacecraft to successfully perform an automated launch, rendezvous, and docking. Today’s reality is that while computer operations are advancing rapidly, science fiction over-simplifies and over-sells current capabilities. This paper discusses the integration of spacecraft computer automation into the operation of one of the United States’ new Commercial Crew vehicles - the Boeing CST-100 Starliner. Lessons learned by the Boeing Mission Operations team, a private-public partnership with NASA, from conceptual design through real-time operation of the first test flight will be discussed. Focus will center on how operations has learned to use the automated software to their advantage while also knowing how to adjust the automation in response to spacecraft or mission anomalies. One goal of advanced spacecraft automation is the ability to reduce both the crew workload and the ground control footprint while at the same time increasing spacecraft and mission flexibility. Historically, crewed spacecraft required a large number of operators on the ground to use a plethora of tools to compute nominal and contingency mission trajectories. Moving those sophisticated software tools to being onboard the vehicle can reduce the need for such complex ground support. Given that today’s spacecraft software is not yet as capable or as flexible in all circumstances as the computers depicted in movies, there is usually a trade-off between software automation cost and the flexibility of that software resulting in a trade-off between what is performed on the spacecraft and what is left to onboard crew or ground control. For missions that go beyond the Moon, software that autonomously controls nearly every aspect of a crewed mission will become a necessity given the long time delays between the spacecraft and Earth’s ground control teams. The lessons learned by Boeing and its Mission Operations team, through the design and implementation of Starliner’s hardware and software automation, will be able to inform future public and private spacecraft design. As the technologies and capabilities evolve, incorporating lessons learned in successful low Earth orbit commercial crew vehicle missions, spacecraft designs will continue to improve and be able to better enable safe execution of human missions to the Moon and beyond.

Robert C Dempsey↗

NASA NTRSDate not supplied

Automated Software for Manned Spacecraft - Bridging the Gap from Sci Fi to Reality

With a voice command or a few taps on the console, the spacecraft pivots on a dime at high velocity and gently docks to an orbiting space platform. This is the image most people have of the complex software computations and integrated hardware performance necessary for a spacecraft to successfully perform an automated launch, rendezvous, and docking. Today’s reality is that while computer operations are advancing rapidly, science fiction over-simplifies and over-sells current capabilities. This paper discusses the integration of spacecraft computer automation into the operation of one of the United States’ new Commercial Crew vehicles - the Boeing CST-100 Starliner. Lessons learned by the Boeing Mission Operations team, a private-public partnership with NASA, from conceptual design through real-time operation of the first test flight will be discussed. Focus will center on how operations has learned to use the automated software to their advantage while also knowing how to adjust the automation in response to spacecraft or mission anomalies. One goal of advanced spacecraft automation is the ability to reduce both the crew workload and the ground control footprint while at the same time increasing spacecraft and mission flexibility. Historically, crewed spacecraft required a large number of operators on the ground to use a plethora of tools to compute nominal and contingency mission trajectories. Moving those sophisticated software tools to being onboard the vehicle can reduce the need for such complex ground support. Given that today’s spacecraft software is not yet as capable or as flexible in all circumstances as the computers depicted in movies, there is usually a trade-off between software automation cost and the flexibility of that software resulting in a trade-off between what is performed on the spacecraft and what is left to onboard crew or ground control. For missions that go beyond the Moon, software that autonomously controls nearly every aspect of a crewed mission will become a necessity given the long time delays between the spacecraft and Earth’s ground control teams. The lessons learned by Boeing and its Mission Operations team, through the design and implementation of Starliner’s hardware and software automation, will be able to inform future public and private spacecraft design. As the technologies and capabilities evolve, incorporating lessons learned in successful low Earth orbit commercial crew vehicle missions, spacecraft designs will continue to improve and be able to better enable safe execution of human missions to the Moon and beyond.

Robert C. Dempsey↗

NASA NTRSDate not supplied

Automated Software for Crewed Spacecraft - Bridging the Gap from Sci Fi to Reality

With a voice command or a few taps on the console, the spacecraft pivots on a dime at high velocity and gently docks to an orbiting space platform. This is the image most people have of the complex software computations and integrated hardware performance necessary for a spacecraft to successfully perform an automated launch, rendezvous, and docking. Today’s reality is that while computer operations are advancing rapidly, science fiction over-simplifies and over-sells current capabilities. This paper discusses the integration of spacecraft computer automation into the operation of one of the United States’ new Commercial Crew vehicles - the Boeing CST-100 Starliner. Lessons learned by the Boeing Mission Operations team, a private-public partnership with NASA, from conceptual design through real-time operation of the first test flight will be discussed. Focus will center on how operations has learned to use the automated software to their advantage while also knowing how to adjust the automation in response to spacecraft or mission anomalies. One goal of advanced spacecraft automation is the ability to reduce both the crew workload and the ground control footprint while at the same time increasing spacecraft and mission flexibility. Historically, crewed spacecraft required a large number of operators on the ground to use a plethora of tools to compute nominal and contingency mission trajectories. Moving those sophisticated software tools to being onboard the vehicle can reduce the need for such complex ground support. Given that today’s spacecraft software is not yet as capable or as flexible in all circumstances as the computers depicted in movies, there is usually a trade-off between software automation cost and the flexibility of that software resulting in a trade-off between what is performed on the spacecraft and what is left to onboard crew or ground control. For missions that go beyond the Moon, software that autonomously controls nearly every aspect of a crewed mission will become a necessity given the long time delays between the spacecraft and Earth’s ground control teams. The lessons learned by Boeing and its Mission Operations team, through the design and implementation of Starliner’s hardware and software automation, will be able to inform future public and private spacecraft design. As the technologies and capabilities evolve, incorporating lessons learned in successful low Earth orbit commercial crew vehicle missions, spacecraft designs will continue to improve and be able to better enable safe execution of human missions to the Moon and beyond.

Robert C Dempsey↗

NASA NTRS1988

Concepts for robot motion primitives required for space station teleoperations

Ground controlled teleoperations are expected to be used to augment Space Station manned extravehicular activities (EVA) and Intravehicular activities (IVA). However, ground controlled teleoperations will encounter communications time delays of from 3 to 8 secs. Time delays greater than 1 sec have been shown to be detrimental to safe and efficient teleoperations. Therefore, concepts must be developed to overcome the hazards and limitations of time delays when performing teleoperations using robots. The concept for robot motion primitives incorporate force/torque and tactile sensor feedback to implement the degree of autonomy required for interactive, ground controlled telerobotics. Several primitives are studied that augment human initiated actions by providing rapid response interaction with the physical environment of a telerobot. These primitives are detailed. They constitute a level of intelligent sensing and reaction required to augment human actions through autonomous interaction with the physical environment.

Grover, Jeffrey L.↗

NASA NTRS2011

Altered Innate and Lymphocytic Immune Responses in Mouse Splenocytes Post-Flight

Space flight is known to affect immune responses of astronauts and animals, decreasing lymphocytic responses to mitogenic stimuli, delayed typed hypersensitivity reactions, and T-cell activation. Despite changes in immune suppression, there are no reports of consistent adverse clinical events post flight. To further investigate the spectrum of affected immune responses, murine splenocytes were stimulated immediately post-shuttle flight (14 days on STS-135) with T-cell stimulators or toll-like receptor agonists. Comparisons were made to ground control splenocytes from age-matched mice. Cell phenotypes were assessed, as well as activation markers and associated cytokine production. The CD4+ population decreased with no concurrent decrease in CD8+ cells from shuttle mice post flight compared to ground controls. Regarding antigen presenting cell populations, the number of CD11c+ cells were slightly elevated post flight, compared to ground controls, with increased MHC Class I expression (I-A(sup b)) and no change in Class II expression (H-2K(sup b)). CD86+ populations were also significantly diminished. However, the decreased markers did not correlate with activity. Stimulation of splenocytes post flight showed significant increase in bead uptake, increased Class I expression, increased TNF-alpha and IL-6 production in response to TLR-2 (zymosan) and TLR-4 (LPS) agonists. While most activated (ConA or anti-CD3/anti-CD28) CD4+ cells showed markedly diminished responses (reduced IL-2 production), non-specific T cell responses to superantigen (SEA/SEB) increased post flight as determined by expression of early activation markers. Production of additional cytokines was also dysregulated postflight. Overall, persistent immune changes during space flight could represent unique clinical risks for exploration class missions. The consequences of pathogenic encounter remain an important concern that should be addressed.

Hwang, ShenAn↗

NASA NTRS2003

Germination and elongation of flax in microgravity

This experiment was conducted as part of a risk mitigation payload aboard the Space Shuttle Atlantis on STS-101. The objectives were to test a newly developed water delivery system, and to determine the optimal combination of water volume and substrate for the imbibition and germination of flax (Linum usitatissimum) seeds in space. Two different combinations of germination paper were tested for their ability to absorb, distribute, and retain water in microgravity. A single layer of thick germination paper was compared with one layer of thin germination paper under a layer of thick paper. Paper strips were cut to fit snugly into seed cassettes, and seeds were glued to them with the micropyle ends pointing outward. Water was delivered in small increments that traveled through the paper via capillary action. Three water delivery volumes were tested, with the largest (480 microliters) outperforming the 400 microliters and 320 microliters volumes for percent germination (90.6%) and root growth (mean=4.1 mm) during the 34-hour spaceflight experiment. The ground control experiment yielded similar results, but with lower rates of germination (84.4%) and shorter root lengths (mean=2.8 mm). It is not clear if the roots emerged more quickly in microgravity and/or grew faster than the ground controls. The single layer of thick germination paper generally exhibited better overall growth than the two layered option. Significant seed position effects were observed in both the flight and ground control experiments. Overall, the design of the water delivery system, seed cassettes and the germination paper strip concept was validated as an effective method for promoting seed germination and root growth under microgravity conditions. c2003 COSPAR. Published by Elsevier Ltd. All rights reserved.

Non-NASA Center↗

NASA NTRS1998

Changes in Arabidopsis leaf ultrastructure, chlorophyll and carbohydrate content during spaceflight depend on ventilation

Leaf structure and function under spaceflight conditions have received little study despite their important implications for biological life support systems using plants. Previous reports described disruption of the membrane apparatus for photosynthesis and a general decrease in carbohydrate content in foliage. During a series of three short-duration experiments (Chromex-03, -04, -05) on the US space shuttle (STS-54, STS-51, STS-68), we examined Arabidopsis thaliana leaves. The plants were at the rosette stage at the time of loading onto the space shuttle, and received the same light, temperature, carbon dioxide and humidity regimes in the orbiter as in ground controls. The experiments differed according to the regime provided in the headspace around the plants: this was either sealed (on mission STS-54); sealed with high levels of carbon dioxide (on mission STS-51) or vented to the cabin air through a filtration system (on mission STS-68). Immediately post-flight, leaf materials were fixed for microscopy or frozen in liquid nitrogen for subsequent analyses of chlorophyll and foliar carbohydrates. At the ultrastructural level, no aberrations in membrane structure were observed in any of the experiments. When air-flow was provided, plastids developed large starch grains in both spaceflight and ground controls. In the experiments with sealed chambers, spaceflight plants differed from ground controls with regard to measured concentrations of carbohydrate and chlorophyll, but the addition of airflow eliminated these differences. The results point to the crucial importance of consideration of the foliage microenvironment when spaceflight effects on leaf structure and metabolism are studied.

STS-51 Shuttle Project↗

NASA NTRS2005

Nutritional and Flavor Components of Brassica Xapa L. Grown on ISS

Brassica rapa L. cv. 'Astroplants' were grown on the International Space Station during April - June 2002 in the Biomass Production System. Plants were manually pollinated and were maturing seeds when they were harvested for preservation in flight by fixation or freezing. Overall growth and development were comparable between flight and ground control plants. Chlorophyll and carbohydrate content of the leaves were the same in the two treatments. Although comparable numbers of seeds were produced inside the seed pods, the developing seeds from the spaceflight treatment had only half of the dry weight of the ground controls and had altered storage components. Glucosinolate content of the stem tissue was also determined. The concentration of 3-butenyl-glucosinolate was on average 75% greater in the spaceflight samples than in the ground control. The results demonstrate how the spaceflight environment influences nutritional and flavor characteristics of a potential crop for use in a Biological Life Support System.

Musgrave, M. E.↗

NASA NTRS1965

Computer-controlled aerospace ground equipment for the aoso communications and data- handling subsystem.

Computer controlled checkout equipment and data handling system for AOSO

GROUND SUPPORT SYSTEM↗

NASA NTRS1999

Tissue Engineered Skeletal Myofibers can Directly "Sense" Gravitational Force Changes

Long-term manned space flight requires a better understanding of skeletal muscle atrophy resulting from microgravity. Atrophy most likely results from changes at both the systemic level (e.g. decreased circulating growth hormone, increased circulating glucocorticoids) and locally (e.g. decreased myofiber resting tension). Differentiated skeletal myofibers in tissue culture have provided a model system over the last decade for gaining a better understanding of the interactions of exogenous growth factors, endogenous growth factors, and muscle fiber tension in regulating protein turnover rates and muscle cell growth. Tissue engineering these cells into three dimensional bioartificial muscle (BAM) constructs has allowed us to extend their use to Space flight studies for the potential future development of countermeasures. Embryonic avian muscle cells were isolated and BAMs tissue engineered as described previously. The myoblasts proliferate and fuse into aligned postmitotic myofibers after ten to fourteen days in vitro. A cylindrical muscle-like structure containing several thousand myofibers is formed which is approximately 30 mm in length, 2-3 mm in diameter, and attached at each end. For the Space Shuttle experiments, the BAMs were transferred to 55 mL bioreactor cartridges (6 BAMs/cartridge). At Kennedy Space Center, the cartridges were mounted in two Space Tissue Loss (STL) Modules (three to four cartridges per Module) and either maintained as ground controls or loaded in a Mid-Deck locker of the Space Shuttle. The BAM cartridges were continuously perfused during the experiment at 1.5 mL/ min with tissue culture medium. Eighteen BAMs were flown for nine days on Mission STS66 while eighteen BAMs served as ground controls. The complete experiment was repeated on Mission STS77 with twenty four BAMs in each group. BAMs could be maintained in a healthy state for at least 30 days in the perfusion bioreactor cartridges. The BAM muscle fibers directly detected both the loss of gravity and the reloading effects of 1 x g. While total cellular metabolism and total protein degradation rates were not altered during 9 to 10 days in Space, protein synthesis rates were significantly reduced and resulted in significant myofiber atrophy compared to ground controls. One g reloading of the flight muscle cells post-flight significantly increased protein synthesis rates and the synthesis rates of myosin heavy chain, fibronectin, and collagen. Tissue cultured muscle cells can directly "sense" changes in gravity and provide a valid model to begin the study of countermeasures. Based on our ground based experiments, and the experiments of others, growth hormone and/or insulin-like growth factors are attractive protein therapeutics which may assisting attenuating skeletal muscle wasting in space. Our laboratory is developing a new cell-based delivery system for this and other potential therapeutic factors for attenuating muscle and bone wasting.

Vandenburgh, Herman H.↗

NASA NTRS2013

Human Machine Interface Programming and Testing

Human Machine Interface (HMI) Programming and Testing is about creating graphical displays to mimic mission critical ground control systems in order to provide NASA engineers with the ability to monitor the health management of these systems in real time. The Health Management System (HMS) is an online interactive human machine interface system that monitors all Kennedy Ground Control Subsystem (KGCS) hardware in the field. The Health Management System is essential to NASA engineers because it allows remote control and monitoring of the health management systems of all the Programmable Logic Controllers (PLC) and associated field devices. KGCS will have equipment installed at the launch pad, Vehicle Assembly Building, Mobile Launcher, as well as the Multi-Purpose Processing Facility. I am designing graphical displays to monitor and control new modules that will be integrated into the HMS. The design of the display screen will closely mimic the appearance and functionality of the actual modules. There are many different field devices used to monitor health management and each device has its own unique set of health management related data, therefore each display must also have its own unique way to display this data. Once the displays are created, the RSLogix5000 application is used to write software that maps all the required data read from the hardware to the graphical display. Once this data is mapped to its corresponding display item, the graphical display and hardware device will be connected through the same network in order to test all possible scenarios and types of data the graphical display was designed to receive. Test Procedures will be written to thoroughly test out the displays and ensure that they are working correctly before being deployed to the field. Additionally, the Kennedy Ground Controls Subsystem's user manual will be updated to explain to the NASA engineers how to use the new module displays.

Foster, Thomas Garrison↗

NASA NTRS1993

A ground track control algorithm for the Topographic Mapping Laser Altimeter (TMLA)

The results of an analysis of an algorithm that will provide autonomous onboard orbit control using orbits determined with Global Positioning System (GPS) data. The algorithm uses the GPS data to (1) compute the ground track error relative to a fixed longitude grid, and (2) determine the altitude adjustment required to correct the longitude error. A program was written on a personal computer (PC) to test the concept for numerous altitudes and values of solar flux using a simplified orbit model including only the J sub 2 zonal harmonic and simple orbit decay computations. The algorithm was then implemented in a precision orbit propagation program having a full range of perturbations. The analysis showed that, even with all perturbations (including actual time histories of solar flux variation), the algorithm could effectively control the spacecraft ground track and yield more than 99 percent Earth coverage in the time required to complete one coverage cycle on the fixed grid (220 to 230 days depending on altitude and overlap allowance).

Blaes, V.↗

NASA NTRS2023

Exploring the ARNTL and RORc Genes' Impact on Lipid Accumulation in the Liver via the Circadian Rhythm

The objective of our proposed experiment is to prove the significance of the upregulation of the ARNTL gene and down regulation of the RORc gene in space, highlighting their crucial roles in causing circadian rhythm misalignment during spaceflight, ultimately resulting in liver lipid accumulation. This experiment was created by thoroughly analyzing the genetic data (UseGalaxy) in OSD-245, Transcriptional analysis of liver from mice flown on the RR-6 mission (NASA GeneLab), and making a comparison between the Flight ISS Terminal and the Ground Control ISS Terminal groups. In our study, we intend to establish experimental groups of Mus musculus (lab mice), with 4 of these groups designated for space travel, while the remaining groups will serve as ground control samples on Earth. Both the space and ground conditions are categorized into four groups according to the genetic makeup of the mice. Within each condition, there will be a control group representing the wild type genes, and the experimental groups will involve knockout mice for ARNTL and RORc. We will then compare the lipid accumulation between the ground control mice and FLT mice to understand whether or not spaceflight causes lipid accumulation, due to a misaligned circadian rhythm. The experiment is anticipated to yield two main outcomes. Firstly, the upregulation of ARNTL is expected to rectify the circadian rhythm misalignment in space, consequently heightening lipid accumulation within the liver. Secondly, the downregulation of RORc is predicted to result in a diminished immune response in the liver, thereby facilitating lipid accumulation. Our space research is expected to aid in lipid buildup countermeasures, informing scientists regarding astronaut protection and providing insights into obesity and non-alcoholic fatty liver disease.

GL4HS↗

NASA NTRSDate not supplied

Transcriptomics-based Machine Learning Analysis Predicts Space-Exposed Murine Livers

Limited sample sizes, high data dimensionality, and sensitivity to technical and biological variability of next generation sequencing (NGS), typically limits machine learning (ML) approaches in spaceflight studies that include radiation effects. However, pooling smaller studies while addressing intra- and inter-study variabilities allows for ML predictive modeling. Here, integration methods were applied to whole transcriptome shotgun sequencing (RNA-seq) data from six mouse liver GeneLab datasets (GLDS) (n ranging from 6 to 39 samples) from with a total of 81 spaceflight and ground-control samples to determine top features (i.e. genes) relevant to spaceflight including the effect of radiation exposure. RNASeq counts were normalized for each study, then merged and scaled across all datasets. Data dimensionality was reduced using a minimum redundancy maximum relevance (MRMR) methodology. Redundancy and relevance were computed using the Pearson correlation and F-statistic, respectively. The top 100 MRMR features were used to predict spaceflight vs. ground-control samples using Random Forest (RF), Support Vector Machine (SVM), and Linear Discriminant Analysis (LDA) classifiers with 5-fold cross validation (CV). Principal component analysis (PCA) on the complete feature set versus the MRMR features shows separation between spaceflight samples and ground controls (Figure 1A). The ML-based gene sets were compared against differential gene expression results obtained with DESeq2 from individual GLDS. Using all features or randomly sampled subsets at matching set sizes with MRMR, a maximum classifier accuracy of 69% was shown on the test set over 5 folds. For all classifiers, CV training using at least the top 30 MRMR genes show minimum 89% accuracy and 0.95 AUC value on the test set over 5 folds (Figure 1B). Baseline set analysis on differentially expressed genes (DEGs) identified using padj ≤ 0.05 show 295 DEGs that overlap at least two studies and 13 DEGs that overlap three studies (Figure 1C). Set analysis between the top 100 MRMR features and the DEGs showed 47 genes that overlap at least one study and 24 genes that overlap two studies. Over-representation analysis showed overlapping biological processes related to fatty acid and lipid metabolism which may indicate these processes in the response to spaceflight stressors. MRMR feature selection for the selected ML methods improve performance relative to a classifier built on all features or randomly sampled subsets. Permutation feature importance within the decorrelated MRMR features showed concordance in feature ranking between ML methods. A challenge of applying ML methods across heterogeneous NGS data is accounting for signal:noise. Here, signal validation across studies was shown by intersecting sets between top MRMR genes and DEGs from DESeq2 analysis. Non-intersecting sets introduce opportunity to explore genes relevant to differentiating space flight exposed groups and implementing ML methods across existing NGS datasets may overcome sample size limitations.

Machine Learning↗

NASA NTRSDate not supplied

Transcriptomics-based Machine Learning (ML) Analysis Predicts Space-Exposed Murine Livers

Limited sample sizes, high data dimensionality, and sensitivity to technical and biological variability of next generation sequencing (NGS), typically limits machine learning (ML) approaches in spaceflight studies that include radiation effects. However, pooling smaller studies while addressing intra- and inter-study variabilities allows for ML predictive modeling. Here, integration methods were applied to whole transcriptome shotgun sequencing (RNA-seq) data from six mouse liver GeneLab datasets (GLDS) (n ranging from 6 to 39 samples) from with a total of 81 spaceflight and ground-control samples to determine top features (i.e. genes) relevant to spaceflight including the effect of radiation exposure. RNASeq counts were normalized for each study, then merged and scaled across all datasets. Data dimensionality was reduced using a minimum redundancy maximum relevance (MRMR) methodology. Redundancy and relevance were computed using the Pearson correlation and F-statistic, respectively. The top 100 MRMR features were used to predict spaceflight vs. ground-control samples using Random Forest (RF), Support Vector Machine (SVM), and Linear Discriminant Analysis (LDA) classifiers with 5-fold cross validation (CV). Principal component analysis (PCA) on the complete feature set versus the MRMR features shows separation between spaceflight samples and ground controls (Figure 1A). The ML-based gene sets were compared against differential gene expression results obtained with DESeq2 from individual GLDS. Using all features or randomly sampled subsets at matching set sizes with MRMR, a maximum classifier accuracy of 69% on the test set over 5 folds. For all classifiers, CV training using at least the top 30 MRMR genes show minimum 89% accuracy and 0.95 AUC value on the test set over 5 folds (Figure 1B). Baseline set analysis on differentially expressed genes (DEGs) identified using padj ≤ 0.05 show 295 DEGs that overlap at least two studies and 13 DEGs that overlap three studies (Figure 1C). Set analysis between the top 100 MRMR features and the DEGs showed 47 genes that overlap at least one study and 24 genes that overlap two studies. Over-representation analysis showed overlapping biological processes related to fatty acid and lipid metabolism which may indicate these processes in the response to spaceflight stressors. MRMR feature selection for the selected ML methods improve performance relative to a classifier built on all features or randomly sampled subsets. Permutation feature importance within the decorrelated MRMR features showed concordance in feature ranking between ML methods. A challenge of applying ML methods across heterogeneous NGS data is accounting for signal:noise. Here, signal validation across studies was shown by intersecting sets between top MRMR genes and DEGs from DESeq2 analysis. Non-intersecting sets introduce opportunity to explore genes relevant to differentiating space flight exposed groups and implementing ML methods across existing NGS datasets may overcome sample size limitations.

Machine Learning↗

NASA NTRSDate not supplied

Transcriptomics-based Machine Learning Analysis Predicts Space-Exposed Murine Livers

Limited sample sizes, high data dimensionality, and sensitivity to technical and biological variability of next generation sequencing (NGS), typically limits machine learning (ML) approaches in spaceflight studies that include radiation effects. However, pooling smaller studies while addressing intra- and inter-study variabilities allows for ML predictive modeling. Here, integration methods were applied to whole transcriptome shotgun sequencing (RNA-seq) data from six mouse liver GeneLab datasets (GLDS) (n ranging from 6 to 39 samples) from with a total of 81 spaceflight and ground-control samples to determine top features (i.e. genes) relevant to spaceflight including the effect of radiation exposure. RNASeq counts were normalized for each study, then merged and scaled across all datasets. Data dimensionality was reduced using a minimum redundancy maximum relevance (MRMR) methodology. Redundancy and relevance were computed using the Pearson correlation and F-statistic, respectively. The top 100 MRMR features were used to predict spaceflight vs. ground-control samples using Random Forest (RF), Support Vector Machine (SVM), and Linear Discriminant Analysis (LDA) classifiers with 5-fold cross validation (CV). Principal component analysis (PCA) on the complete feature set versus the MRMR features shows separation between spaceflight samples and ground controls (Figure 1A). The ML-based gene sets were compared against differential gene expression results obtained with DESeq2 from individual GLDS. Using all features or randomly sampled subsets at matching set sizes with MRMR, a maximum classifier accuracy of 69% was shown on the test set over 5 folds. For all classifiers, CV training using at least the top 30 MRMR genes show minimum 89% accuracy and 0.95 AUC value on the test set over 5 folds (Figure 1B). Baseline set analysis on differentially expressed genes (DEGs) identified using padj ≤ 0.05 show 295 DEGs that overlap at least two studies and 13 DEGs that overlap three studies (Figure 1C). Set analysis between the top 100 MRMR features and the DEGs showed 47 genes that overlap at least one study and 24 genes that overlap two studies. Over-representation analysis showed overlapping biological processes related to fatty acid and lipid metabolism which may indicate these processes in the response to spaceflight stressors. MRMR feature selection for the selected ML methods improve performance relative to a classifier built on all features or randomly sampled subsets. Permutation feature importance within the decorrelated MRMR features showed concordance in feature ranking between ML methods. A challenge of applying ML methods across heterogeneous NGS data is accounting for signal:noise. Here, signal validation across studies was shown by intersecting sets between top MRMR genes and DEGs from DESeq2 analysis. Non-intersecting sets introduce opportunity to explore genes relevant to differentiating space flight exposed groups and implementing ML methods across existing NGS datasets may overcome sample size limitations.

Machine Learning↗