Search NASA⌕ Search

SEARCH · Search NASA

Results for “Gene expression data”

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 1,261 records · Page 70

The Effect of Spaceflight on Bone Cell Cultures

Understanding the response of bone to mechanical loading (unloading) is extremely important in defining the means of adaptation of the body to a variety of environmental conditions such as during heightened physical activity or in extended explorations of space or the sea floor. The mechanisms of the adaptive response of bone are not well defined, but undoubtedly they involve changes occurring at the cellular level of bone structure. This proposal has intended to examine the hypothesis that the loading (unloading) response of bone is mediated by specific cells through modifications of their activity cytoskeletal elements, and/or elaboration of their extracellular matrices. For this purpose, this laboratory has utilized the results of a number of previous studies defining molecular biological, biochemical, morphological, and ultrastructural events of the reproducible mineralization of a primary bone cell (osteoblast) culture system under normal loading (1G gravity level). These data and the culture system then were examined following the use of the cultures in two NASA shuttle flights, STS-59 and STS-63. The cells collected from each of the flights were compared to respective synchronous ground (1G) control cells examined as the flight samples were simultaneously analyzed and to other control cells maintained at 1G until the time of shuttle launch, at which point they were terminated and studied (defined as basal cells). Each of the cell cultures was assayed in terms of metabolic markers- gene expression; synthesis and secretion of collagen and non-collagenous proteins, including certain cytoskeletal components; assembly of collagen into macrostructural arrays- formation of mineral; and interaction of collagen and mineral crystals during calcification of the cultures. The work has utilized a combination of biochemical techniques (radiolabeling, electrophoresis, fluorography, Western and Northern Blotting, and light microscopic immunofluorescence) and structural methods (conventional and high voltage electron microscopy, inununocytochemistry, stereomicroscopy, and 3D image reconstruction). The studies have provided new knowledge of aspects of bone cell development and structural regulation, extracellular matrix assembly, and mineralization during spaceflight and under normal gravity. The information has contributed to insights into the means in general by which cells respond and adapt to different conditions of gravity (loading). The data may as well have suggested an underlying basis for the observed loss of bone by vertebrates, including man, in microgravity; and these scientific results may have implications for understanding bone loss following fracture healing and extended periods of inactivity such as during long-term bedrest.

Landis, William J.↗

Bystander Effects of Ionizing Radiation

The objectives of this grant renewal are to provide administrative support and travel funds to allow the continued participation of the principal investigator (Dr. John B. Little) as an advisor to research initiated by several research fellows from his laboratory. The actual research will be carried out under the direction of Dr. Hatsumi Nagasawa with the collaboration of Dr. Joel Bedford at the Colorado State University, and by Drs. Edouard Azzam and Sonia de Toledo at the University of Medicine and Dentistry of New Jersey. Dr. Little will advise on the planning of experiments and development of experimental protocols, the analysis of data, and the preparation of manuscripts for publication. The Specific Aims for several of the planned experiments include: 1) to extend studies of the role of recombinational repair in the bystander effect by examining other genes in this pathway and cell lines deficient in excision repair; 2) to continue studies to determine the nature of the damage signal transmitted to bystander cells including the expression of several connexins in the bystander response, and the extent to which the enhanced oxidative metabolism observed in bystander cells may relate to the nature of the transmitted bystander signal; 3) to utilize a genome-wide approach to examine the genetic basis for the hypersensitivity to ionization we have observed in unaffected parents of patients with hereditary retinoblastoma, as well as from a group of apparently normal individuals that show similar radiosensitivity; 4) to complete studies concerning the induction of high frequencies of cells with massive chromosome damage in clonal derivatives of p53 and p21 knockout mouse cell lines; in particular to examine the role of telomere changes in this phenomenon. Overall, the results of these studies should enhance our understanding of the risk of low-dose exposures to ionizing radiation, including human populations to residential radon as well as occupational exposures.

63 RADIATION, THERMAL, AND OTHER ENVIRON. POLLUTAN↗

X-ray structure of putative acyl-ACP desaturase DesA2 from Mycobacterium tuberculosis H37Rv

Genome sequencing showed that two proteins in Mycobacterium tuberculosis H37Rv contain the metal binding motif (D/E)X{sub 2}HX{sub {approx}100}(D/E)X{sub 2}H characteristic of the soluble diiron enzyme superfamily. These putative acyl-ACP desaturase genes desA1 and desA2 were cloned from genomic DNA and expressed in Escherichia coli BL21(DE3). DesA1 was found to be insoluble, but in contrast, DesA2 was a soluble protein amenable to biophysical characterization. Here, we report the 2.0 {angstrom} resolution X-ray structure of DesA2 determined by multiple anomalous dispersion (MAD) phasing from a Se-met derivative and refinement against diffraction data obtained on the native protein. The X-ray structure shows that DesA2 is a homodimeric protein with a four-helix bundle core flanked by five additional helices that overlay with 192 structurally equivalent amino acids in the structure of stearoyl-ACP {Delta}9 desaturase from castor plant with an rms difference 1.42 {angstrom}. In the DesA2 crystals, one metal (likely Mn from the crystallization buffer) was bound in high occupancy at the B-site of the conserved metal binding motif, while the A-site was not occupied by a metal ion. Instead, the amino group of Lys-76 occupied this position. The relationships between DesA2 and known diiron enzymes are discussed.

59 BASIC BIOLOGICAL SCIENCES↗

Activation of peroxisome proliferator-activated receptor-{alpha} enhances fatty acid oxidation in human adipocytes

Highlights: {yields} PPAR{alpha} activation increased mRNA expression levels of adipocyte differentiation marker genes and GPDH activity in human adipocytes. {yields} PPAR{alpha} activation also increased insulin-dependent glucose uptake in human adipocytes. {yields} PPAR{alpha} activation did not affect lipid accumulation in human adipocytes. {yields} PPAR{alpha} activation increased fatty acid oxidation through induction of fatty acid oxidation-related genes in human adipocytes. -- Abstract: Peroxisome proliferator-activated receptor-{alpha} (PPAR{alpha}) is a key regulator for maintaining whole-body energy balance. However, the physiological functions of PPAR{alpha} in adipocytes have been unclarified. We examined the functions of PPAR{alpha} using human multipotent adipose tissue-derived stem cells as a human adipocyte model. Activation of PPAR{alpha} by GW7647, a potent PPAR{alpha} agonist, increased the mRNA expression levels of adipocyte differentiation marker genes such as PPAR{gamma}, adipocyte-specific fatty acid-binding protein, and lipoprotein lipase and increased both GPDH activity and insulin-dependent glucose uptake level. The findings indicate that PPAR{alpha} activation stimulates adipocyte differentiation. However, lipid accumulation was not changed, which is usually observed when PPAR{gamma} is activated. On the other hand, PPAR{alpha} activation by GW7647 treatment induced the mRNA expression of fatty acid oxidation-related genes such as CPT-1B and AOX in a PPAR{alpha}-dependent manner. Moreover, PPAR{alpha} activation increased the production of CO{sub 2} and acid soluble metabolites, which are products of fatty acid oxidation, and increased oxygen consumption rate in human adipocytes. The data indicate that activation of PPAR{alpha} stimulates both adipocyte differentiation and fatty acid oxidation in human adipocytes, suggesting that PPAR{alpha} agonists could improve insulin resistance without lipid accumulation in adipocytes. The expected effects of PPAR{alpha} activation are very valuable for managing diabetic conditions accompanied by obesity, because PPAR{gamma} agonists, usually used as antidiabetic drugs, induce excessive lipid accumulation in adipocytes in addition to improvement of insulin resistance.

60 APPLIED LIFE SCIENCES↗

Evolutionary characteristics of intergenic transcribed regions indicate rare novel genes and widespread noisy transcription in the Poaceae

Abstract Extensive transcriptional activity occurring in intergenic regions of genomes has raised the question whether intergenic transcription represents the activity of novel genes or noisy expression. To address this, we evaluated cross-species and post-duplication sequence and expression conservation of intergenic transcribed regions (ITRs) in four Poaceae species. Among 43,301 ITRs across the four species, 34,460 (80%) are species-specific. ITRs found across species tend to be more divergent in expression and have more recent duplicates compared to annotated genes. To assess if ITRs are functional (under selection), machine learning models were established in Oryza sativa (rice) that could accurately distinguish between phenotype genes and pseudogenes (area under curve-receiver operating characteristic = 0.94). Based on the models, 584 (8%) and 4391 (61%) rice ITRs are classified as likely functional and nonfunctional with high confidence, respectively. ITRs with conserved expression and ancient retained duplicates, features that were not part of the model, are frequently classified as likely-functional, suggesting these characteristics could serve as pragmatic rules of thumb for identifying candidate sequences likely to be under selection. This study also provides a framework to identify novel genes using comparative transcriptomic data to improve genome annotation that is fundamental for connecting genotype to phenotype in crop and model systems.

59 BASIC BIOLOGICAL SCIENCES↗

Molecular cloning of a novel nuclear factor, TDRP1, in spermatogenic cells of testis and its relationship with spermatogenesis

We reported the identification of a novel gene termed TDRP (encoding testis development-related protein) that might be involved in spermatogenesis. The human TDRP gene had two distinct transcripts, TDRP1 and TDRP2, which encoded proteins of 183 aa and 198 aa respectively. Tdrp mRNA was predominantly expressed in testis tissue. We generated rabbit polyclonal antibodies specific against human TDRP1. Immunohistochemistry analysis showed TDRP1 was expressed in spermatogenic cells, especially with high expression in spermatocytes. We provided evidence that TDRP1 distributed in both cytoplasm and nuclei of spermatogenic cells. Expression patterns of Tdrp1 mRNA and its protein were investigated in the rat testis tissues of different developmental stages. Both Tdrp1 mRNA and its protein were barely detected in the testis of neonatal rats, increased remarkably at 3 weeks postpartum, and peaked at 2 months postpartum. We also investigated TDRP1 expressions in testis tissues of azoospermic men with defective spermatogenesis. Western blot analysis showed that TDRP1 expressions were significantly lower in the testis tissues of azoospermic men compared with normal controls. These current data demonstrated that as a nuclear factor, TDRP1 might play an important role in spermatogenesis.

60 APPLIED LIFE SCIENCES↗

Identification of Small RNAs in Desulfovibrio vulgaris Hildenborough

Desulfovibrio vulgaris is an anaerobic sulfate-reducing bacterium capable of facilitating the removal of toxic metals such as uranium from contaminated sites via reduction. As such, it is essential to understand the intricate regulatory cascades involved in how D. vulgaris and its relatives respond to stressors in such sites. One approach is the identification and analysis of small non-coding RNAs (sRNAs); molecules ranging in size from 20-200 nucleotides that predominantly affect gene regulation by binding to complementary mRNA in an anti-sense fashion and therefore provide an immediate regulatory response. To identify sRNAs in D. vulgaris, a bacterium that does not possess an annotated hfq gene, RNA was pooled from stationary and exponential phases, nitrate exposure, and biofilm conditions. The subsequent RNA was size fractionated, modified, and converted to cDNA for high throughput transcriptomic deep sequencing. A computational approach to identify sRNAs via the alignment of seven separate Desulfovibrio genomes was also performed. From the deep sequencing analysis, 2,296 reads between 20 and 250 nt were identified with expression above genome background. Analysis of those reads limited the number of candidates to ~;;87 intergenic, while ~;;140 appeared to be antisense to annotated open reading frames (ORFs). Further BLAST analysis of the intergenic candidates and other Desulfovibrio genomes indicated that eight candidates were likely portions of ORFs not previously annotated in the D. vulgaris genome. Comparison of the intergenic and antisense data sets to the bioinformatical predicted candidates, resulted in ~;;54 common candidates. Current approaches using Northern analysis and qRT-PCR are being used toverify expression of the candidates and to further develop the role these sRNAs play in D. vulgaris regulation.

54 ENVIRONMENTAL SCIENCES↗

Quantitative multiplexed analysis of gene and protein expression patterns in Yarrowia lipolytica

In this report, we present coordinated observations of protein and mRNA transcript counts at the single-cell level in the oleaginous yeast model Yarrowia lipolytica. The transcription factor Xbp1p regulates entry into a quiescent state, representing a shift of resources to sequestration of nutrients rather than cell division. We observed the responses of wild-type and Δxbp1 cells to protein (by fluorescence) and transcript quantification and localization at both single-cell and population-averaged levels. Data were collected via single-molecule fluorescence in situ hybridization (smFISH) and qPCR under nitrogen depletion, a condition that drives lipid accumulation. These techniques reveal a complex and heterogeneous population of Xbp1p dynamics and downstream regulation. Our findings highlight the need for single-cell resolution analyses to describe cellular dynamics and regulatory processes.

Yarrowia lipolytica↗

Comparative genomic and transcriptomic analyses of trans-kingdom pathogen Fusarium solani species complex reveal degrees of compartmentalization

Background: The Fusarium solani species complex (FSSC) comprises fungal pathogens responsible for mortality in a diverse range of animals and plants, but their genome diversity and transcriptome responses in animal pathogenicity remain to be elucidated. We sequenced, assembled and annotated six chromosome-level FSSC clade 3 genomes of aquatic animal and plant host origins. We established a pathosystem and investigated the expression data of F. falciforme and F. keratoplasticum in Chinese softshell turtle (Pelodiscus sinensis) host. Results: Comparative analyses between the FSSC genomes revealed a spectrum of conservation patterns in chromosomes categorised into three compartments: core, fast-core (FC), and lineage-specific (LS). LS chromosomes contribute to variations in genomes size, with up to 42.2% of variations between F. vanettenii strains. Each chromosome compartment varied in structural architectures, with FC and LS chromosomes contain higher proportions of repetitive elements with genes enriched in functions related to pathogenicity and niche expansion. We identified differences in both selection in the coding sequences and DNA methylation levels between genome features and chromosome compartments which suggest a multi-speed evolution that can be traced back to the last common ancestor of Fusarium. We further demonstrated that F. falciforme and F. keratoplasticum are opportunistic pathogens by inoculating P. sinensis eggs and identified differentially expressed genes also associated with plant pathogenicity. These included the most upregulated genes encoding the CFEM (Common in Fungal Extracellular Membrane) domain. Conclusions: The high-quality genome assemblies provided new insights into the evolution of FSSC chromosomes, which also serve as a resource for studies of fungal genome evolution and pathogenesis. This study also establishes an animal model for fungal pathogens of trans-kingdom hosts.

59 BASIC BIOLOGICAL SCIENCES↗

Regulation of the pro-inflammatory cytokine osteopontin by GIP in adipocytes - A role for the transcription factor NFAT and phosphodiesterase 3B

Highlights: Black-Right-Pointing-Pointer GIP stimulates lipogenesis and osteopontin expression in primary adipocytes. Black-Right-Pointing-Pointer GIP-induced osteopontin expression is NFAT-dependent. Black-Right-Pointing-Pointer Osteopontin expression is PDE3-dependent. Black-Right-Pointing-Pointer Osteopontin expression is increased in PDE3B KO mice. -- Abstract: The incretin - glucose-dependent insulinotropic polypeptide (GIP) - and the pro-inflammatory cytokine osteopontin are known to have important roles in the regulation of adipose tissue functions. In this work we show that GIP stimulates lipogenesis and osteopontin expression in primary adipocytes. The GIP-induced increase in osteopontin expression was inhibited by the NFAT (the transcription factor nuclear factor of activated T-cells) inhibitor A-285222. Also, the NFAT kinase glycogen synthase kinase (GSK) 3 was upregulated by GIP. To test whether cAMP might be involved in GIP-mediated effects on osteopontin a number of strategies were used. Thus, the {beta}3-adrenergic receptor agonist CL316,243 stimulated osteopontin expression, an effects which was mimicked by OPC3911, a specific inhibitor of phosphodiesterase 3. Furthermore, treatment of phosphodiesterase 3B knock-out mice with CL316,243 resulted in a dramatic upregulation of osteopontin in adipose tissue which was not the case in wild-type mice. In summary, we delineate mechanisms by which GIP stimulates osteopontin in adipocytes. Given the established link between osteopontin and insulin resistance, our data suggest that GIP by stimulating osteopontin expression, also could promote insulin resistance in adipocytes.

60 APPLIED LIFE SCIENCES↗

The mitochondrial fatty acid synthesis (mtFASII) pathway is capable of mediating nuclear-mitochondrial cross talk through the PPAR system of transcriptional activation

Highlights: •The function of the mitochondria fatty acid synthesis pathway is partially unknown. •Overexpression of the pathway causes transcriptional activation through PPARs. •Knock down of the pathway attenuates that activation. •The last enzyme in the pathway regulates its own transcription. •Products of the mtFASII pathway are able to drive nuclear transcription. -- Abstract: Mammalian cells contain two fatty acid synthesis pathways, the cytosolic FASI pathway, and the mitochondrial FASII pathway. The selection behind the conservation of the mitochondrial pathway is not completely understood, given the presence of the cytosolic FAS pathway. In this study, we show through heterologous gene reporter systems and PCR-based arrays that overexpression of MECR, the last step in the mtFASII pathway, causes modulation of gene expression through the PPAR pathway. Electromobility shift assays (EMSAs) demonstrate that overexpression of MECR causes increased binding of PPARs to DNA, while cell fractionation and imaging studies show that MECR remains localized to the mitochondria. Interestingly, knock down of the mtFASII pathway lessens the effect of MECR on this transcriptional modulation. Our data are most consistent with MECR-mediated transcriptional activation through products of the mtFASII pathway, although we cannot rule out MECR acting as a coactivator. Further investigation into the physiological relevance of this communication will be necessary to better understand some of the phenotypic consequences of deficits in this pathway observed in animal models and human disease.

60 APPLIED LIFE SCIENCES↗

Activation of farnesoid X receptor induces RECK expression in mouse liver

Highlights: •RECK is a novel transcriptional target gene of FXR in mouse liver. •The FXR response element is located within the intron 1 of RECK gene. •FXR agonist reverses the down-regulation of RECK in the liver in mouse NASH model. -- Abstract: Farnesoid X receptor (FXR) belongs to the ligand-activated nuclear receptor superfamily, and functions as a transcription factor regulating the transcription of numerous genes involved in bile acid homeostasis, lipoprotein and glucose metabolism. In the present study, we identified RECK, a membrane-anchored inhibitor of matrix metalloproteinases, as a novel target gene of FXR in mouse liver. We found that FXR agonist substantially augmented hepatic RECK mRNA and protein expression in vivo and in vitro. FXR regulated the transcription of RECK through directly binding to FXR response element located within intron 1 of the mouse RECK gene. Moreover, FXR agonist reversed the down-regulation of RECK in the livers from mice fed a methionine and choline deficient diet. In summary, our data suggest that RECK is a novel transcriptional target of FXR in mouse liver, and provide clues to better understanding the function of FXR in liver.

60 APPLIED LIFE SCIENCES↗

Decomprolute is a benchmarking platform designed for multiomics-based tumor deconvolution

Tumor deconvolution is a reliable way to disentangle the diverse cell types that comprise solid tumors. To date, however, both the algorithms developed to deconvolve tumor samples, and the gold standard datasets used to assess the algorithms are geared toward the analysis of gene expression (e.g., RNA-seq) rather than protein levels in tumor cells. While gene expression is less expensive to measure, protein levels provide a more accurate view of immune markers. To facilitate the development as well as improve the reproducibility and reusability of multi-omic deconvolution algorithms, we introduce Decomprolute, a Common Workflow Language framework that leverages containerization to compare tumor deconvolution algorithms across multiomic data sets. Decomprolute incorporates the large-scale multiomic data sets produced by the Clinical Proteomic Tumor Analysis Consortium (CPTAC), which include matched mRNA expression and proteomic data from thousands of tumors across multiple cancer types to build a fully open-source, containerized proteogenomic tumor deconvolution benchmarking platform. The platform consists of modular architecture and it comes with well-defined input and output formats at each module. As a result, it is robust and extendable easily with additional algorithms or analyses. The platform is available for access and use at http://pnnl-compbio.github.io/decomprolute.

60 APPLIED LIFE SCIENCES↗

The origin of aerobic methanotrophy within the Proteobacteria

ABSTRACT Aerobic methanotrophs play critical roles in the global carbon cycle, but despite their environmental ubiquity, they are phylogenetically restricted. Via bioinformatic analyses, it is shown that methanotrophy likely arose from methylotrophy from the lateral gene transfer of either of the two known forms of methane monooxygenase (particulate and soluble methane monooxygenases). Moreover, it appears that both known forms of pyrroloquinoline quinone-dependent methanol dehydrogenase (MeDH) found in methanotrophs—the calcium-containing Mxa-MeDH and the rare earth element-containing Xox-MeDH—were likely encoded in the genomes before the acquisition of the methane monooxygenases (MMOs), but that some methanotrophs subsequently received an additional copy of Xox-MeDH-encoding genes via lateral gene transfer. Further, data are presented that indicate the evolution of methanotrophy from methylotrophy not only required lateral transfer of genes encoding for methane monooxygenases, but also likely the pre-existence of a means of collecting copper. Given the emerging interest in valorizing methane via biological platforms, it is recommended that future strategies for heterologous expression of methane monooxygenase for conversion of methane to methanol also include cloning of genes encoding mechanism(s) of copper uptake, especially for expression of particulate methane monooxygenase.

59 BASIC BIOLOGICAL SCIENCES↗

Glycosyl transferase GT2 genes mediate the biosynthesis of an unusual (1,3;1,4)‐β‐glucan exopolysaccharide in the bacterium Sarcina ventriculi

Abstract Linear, unbranched (1,3;1,4)‐β‐glucans (mixed‐linkage glucans or MLGs) are commonly found in the cell walls of grasses, but have also been detected in basal land plants, algae, fungi and bacteria. Here we show that two family GT2 glycosyltransferases from the Gram‐positive bacteriumSarcina ventriculiare capable of synthesizing MLGs. Immunotransmission electron microscopy demonstrates that MLG is secreted as an exopolysaccharide, where it may play a role in organizing individual cells into packets that are characteristic ofSarcinaspecies. Heterologous expression of these two genes shows that they are capable of producing MLGsin planta, including an MLG that is chemically identical to the MLG secreted fromS. ventriculicells but which has regularly spaced (1,3)‐β‐linkages in a structure not reported previously for MLGs. The tandemly arranged, paralogous pair of genes are designatedSvBmlgs1andSvBmlgs2. The data indicate that MLG synthases have evolved different enzymic mechanisms for the incorporation of (1,3)‐β‐ and (1,4)‐β‐glucosyl residues into a single polysaccharide chain. Amino acid variants associated with the evolutionary switch from (1,4)‐β‐glucan (cellulose) to MLG synthesis have been identified in the active site regions of the enzymes. The presence of MLG synthesis in bacteria could prove valuable for large‐scale production of MLG for medical, food and beverage applications.

Biochemistry & Molecular Biology↗

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↗

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↗