Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “gene expression”

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 379 records · Page 21

Genome-scale models of metabolism and gene expression extend and refine growth phenotype prediction

Growth is a fundamental process of life. Growth requirements are well-characterized experimentally for many microbes; however, we lack a unified model for cellular growth. Such a model must be predictive of events at the molecular scale and capable of explaining the high-level behavior of the cell as a whole. Here, we construct an ME-Model for Escherichia coli—a genome-scale model that seamlessly integrates metabolic and gene product expression pathways. The model computes B80% of the functional proteome (by mass), which is used by the cell to support growth under a given condition. Metabolism and gene expression are interdependent processes that affect and constrain each other. We formalize these constraints and apply the principle of growth optimization to enable the accurate prediction of multi-scale phenotypes, ranging from coarse-grained (growth rate, nutrient uptake, by-product secretion) to fine-grained (metabolic fluxes, gene expression levels). Our results unify many existing principles developed to describe bacterial growth.

59 BASIC BIOLOGICAL SCIENCES↗

Potentiating antibiotic efficacy via perturbation of non-essential gene expression

Proliferation of multidrug-resistant (MDR) bacteria poses a threat to human health, requiring new strategies. Here we propose using fitness neutral gene expression perturbations to potentiate antibiotics. We systematically explored 270 gene knockout-antibiotic combinations in Escherichia coli, identifying 90 synergistic interactions. Identified gene targets were subsequently tested for antibiotic synergy on the transcriptomic level via multiplexed CRISPR-dCas9 and showed successful sensitization of E. coli without a separate fitness cost. These fitness neutral gene perturbations worked as co-therapies in reducing a Salmonella enterica intracellular infection in HeLa. Finally, these results informed the design of four antisense peptide nucleic acid (PNA) co-therapies, csgD, fnr, recA and acrA, against four MDR, clinically isolated bacteria. PNA combined with sub-minimal inhibitory concentrations of trimethoprim against two isolates of Klebsiella pneumoniae and E. coli showed three cases of re-sensitization with minimal fitness impacts. Our results highlight a promising approach for extending the utility of current antibiotics.

60 APPLIED LIFE SCIENCES↗

Design-driven optimization of low-cost reagent formulations for reproducible and high-yielding cell-free gene expression

Access to recombinant proteins is vital in basic science and biotechnology research. Cell-free gene expression systems provide one approach to address this need, but widespread utilization remains limited by the cost, complexity, and inconsistency of current platforms. To address these limitations, we carry out a multi-dimensional definitive screening design to reduce the number of reagent components and remove costly secondary energy substrates. From 1,231 different reagent formulations, we discover a simple and reproducible system based on 12 components. The optimized reagent formulation can produce 2.4 ± 0.3 g/L of protein product at the 15-µL scale (~$\$60$/gprotein) and 3.7 ± 0.2 g/L (~$\$39$/gprotein) at the 4-mL scale with oxygen supplementation. This provides an average 95% reduction in cost over previous cell-free reagent formulations. We further show that the optimized reagent formulation can produce nucleoside triphosphates from nitrogenous bases and ribose and that it is robust to failure across batches of cell lysates, users/locations, and in the synthesis of more than 20 different proteins. For example, we demonstrate the production of fifteen therapeutically relevant products, including full-length aglycosylated monoclonal antibodies. We anticipate that our optimized reagent formulation will democratize the use of cell-free systems for protein manufacturing and synthetic biology applications.

Biologics↗

Exposure to the synthetic FXR agonist GW4064 causes alterations in gene expression and sublethal hepatotoxicity in eleutheroembryo medaka (Oryzias latipes)

The small freshwater teleost, medaka (Oryzias latipes), has a history of usage in studies of chronic toxicity of liver and biliary system. Recent progress with this model has focused on defining the medaka hepatobiliary system. Here we investigate critical liver function and toxicity by examining the in vivo role and function of the farnesoid X receptor alpha (FXRalpha, NR1H4), a member of the nuclear receptor superfamily that plays an essential role in the regulation of bile acid homeostasis. Quantitative mRNA analysis of medaka FXRalpha demonstrates differential expression of two FXRalpha isoforms designated Fxralpha1 and Fxralpha2, in both free swimming medaka embryos with remaining yolk (eleutheroembryos, EEs) and adults. Activation of medaka Fxralpha in vivo with GW4064 (a strong FXRalpha agonist) resulted in modification of gene expression for defined FXRalpha gene targets including the bile salt export protein, small heterodimer partner, and cytochrome P450 7A1. Histological examination of medaka liver subsequent to GW4064 exposure demonstrated significant lipid accumulation, cellular and organelle alterations in both hepatocytes and biliary epithelial cells of the liver. This report of hepatobiliary injury following GW4064 exposure extends previous investigations of the intrahepatic biliary system in medaka, reveals sensitivity to toxicant exposure, and illustrates the need for added resolution in detection and interpretation of toxic responses in this vertebrate.

60 APPLIED LIFE SCIENCES↗

PPAR{alpha} does not suppress muscle-associated gene expression in brown adipocytes but does influence expression of factors that fingerprint the brown adipocyte

Brown adipocytes and myocytes develop from a common adipomyocyte precursor. PPAR{alpha} is a nuclear receptor important for lipid and glucose metabolism. It has been suggested that in brown adipose tissue, PPAR{alpha} represses the expression of muscle-associated genes, in this way potentially acting to determine cell fate in brown adipocytes. To further understand the possible role of PPAR{alpha} in these processes, we measured expression of muscle-associated genes in brown adipose tissue and brown adipocytes from PPAR{alpha}-ablated mice, including structural genes (Mylpf, Tpm2, Myl3 and MyHC), regulatory genes (myogenin, Myf5 and MyoD) and a myomir (miR-206). However, in our hands, the expression of these genes was not influenced by the presence or absence of PPAR{alpha}, nor by the PPAR{alpha} activator Wy-14,643. Similarly, the expression of genes common for mature brown adipocyte and myocytes (Tbx15, Meox2) were not affected. However, the brown adipocyte-specific regulatory genes Zic1, Lhx8 and Prdm16 were affected by PPAR{alpha}. Thus, it would not seem that PPAR{alpha} represses muscle-associated genes, but PPAR{alpha} may still play a role in the regulation of the bifurcation of the adipomyocyte precursor into a brown adipocyte or myocyte phenotype.

60 APPLIED LIFE SCIENCES↗

Effect of the anatomical site on telomere length and pref-1 gene expression in bovine adipose tissues

Adipose tissue growth is associated with preadipocyte proliferation and differentiation. Telomere length is a biological marker for cell proliferation. Preadipocyte factor-1 (pref-1) is specifically expressed in preadipocytes and acts as a molecular gatekeeper of adipogenesis. In the present study, we investigated the fat depot-specific differences in telomere length and pref-1 gene expression in various anatomical sites (subcutaneous, intramuscular and visceral) of fattening Wagyu cattle. Visceral adipose tissue expressed higher pref-1 mRNA than did subcutaneous and intramuscular adipose tissues. The telomere length in visceral adipose tissue tended to be longer than that of subcutaneous and intramuscular adipose tissues. The telomere length of adipose tissue was not associated with adipocyte size from three anatomical sites. No significant correlation was found between the pref-1 mRNA level and the subcutaneous adipocyte size. In contrast, the pref-1 mRNA level was negatively correlated with the intramuscular and visceral adipocyte size. These results suggest that anatomical sites of adipose tissue affect the telomere length and expression pattern of the pref-1 gene in a fat depot-specific manner. - Highlights: • Visceral adipose tissue express higher pref-1 mRNA than other anatomical sites. • Telomere length in visceral adipose tissue is longer than other anatomical sites. • Telomere length of adipose tissue is not associated with adipocyte size. • Pref-1 mRNA is negatively correlated with intramuscular and visceral adipocyte size.

60 APPLIED LIFE SCIENCES↗

Gene expression for biodosimetry and effect prediction purposes: promises, pitfalls and future directions – key session ConRad 2021

In a nuclear or radiological event, an early diagnostic or prognostic tool is needed to distinguish unexposed from low- and highly exposed individuals with the latter requiring early and intensive medical care. Radiation-induced gene expression (GE) changes observed within hours and days after irradiation have shown potential to serve as biomarkers for either dose reconstruction (retrospective dosimetry) or the prediction of consecutively occurring acute or chronic health effects. The advantage of GE markers lies in their capability for early (1–3 days after irradiation), high-throughput, and point-of-care (POC) diagnosis required for the prediction of the acute radiation syndrome (ARS).

61 RADIATION PROTECTION AND DOSIMETRY↗

Investigating the Role of TNF-α and IFN-γ Activation on the Dynamics of iNOS Gene Expression in LPS Stimulated Macrophages

Macrophage produced inducible nitric oxide synthase (iNOS) is known to play a critical role in the proinflammatory response against intracellular pathogens by promoting the generation of bactericidal reactive nitrogen species. Robust and timely production of nitric oxide (NO) by iNOS and analogous production of reactive oxygen species are critical components of an effective immune response. In addition to pathogen associated lipopolysaccharides (LPS), iNOS gene expression is dependent on numerous proinflammatory cytokines in the cellular microenvironment of the macrophage, two of which include interferon gamma (IFN-γ) and tumor necrosis factor alpha (TNF-α). To understand the synergistic effect of IFN-γ and TNF-α activation, and LPS stimulation on iNOS expression dynamics and NO production, we developed a systems biology based mathematical model. Using our model, we investigated the impact of pre-infection cytokine exposure, or priming, on the system. We explored the essentiality of IFN-γ priming to the robustness of initial proinflammatory response with respect to the ability of macrophages to produce reactive species needed for pathogen clearance. Results from our theoretical studies indicated that IFN-γ and subsequent activation of IRF1 are essential in consequential production of iNOS upon LPS stimulation. We showed that IFN-γ priming at low concentrations greatly increases the effector response of macrophages against intracellular pathogens. Ultimately the model demonstrated that although TNF-α contributed towards a more rapid response time, measured as time to reach maximum iNOS production, IFN-γ stimulation was significantly more significant in terms of the maximum expression of iNOS and the concentration of NO produced.

59 BASIC BIOLOGICAL SCIENCES↗

Structural analysis of the NifL-NifA complex reveals the molecular basis of anti-activation of nitrogen fixation gene expression in Azotobacter vinelandii

Understanding the molecular basis of regulated nitrogen (N 2 ) fixation is essential for engineering N 2 -fixing bacteria that fulfill the demand of crop plants for fixed nitrogen, reducing our reliance on synthetic nitrogen fertilizers. In Azotobacter vinelandii and many other members of Proteobacteria, the two-component system comprising the anti-activator protein (NifL) and the Nif-specific transcriptional activator (NifA)controls the expression of nif genes, encoding the nitrogen fixation machinery. The NifL-NifA system evolved the ability to integrate several environmental cues, such as oxygen, nitrogen, and carbon availability. The nitrogen fixation machinery is thereby only activated under strictly favorable conditions, enabling diazotrophs to thrive in competitive environments. While genetic and biochemical studies have enlightened our understanding of how NifL represses NifA, the molecular basis of NifA sequestration by NifL depends on structural information on their interaction. Here, we present mechanistic insights into how nitrogen fixation is regulated by combining biochemical and genetic approaches with a low-resolution cryo-electron microscopy (cryo-EM) map of the oxidized NifL-NifA complex. Our findings define the interaction surface between NifL and NifA and reveal how this interaction can be manipulated to generate bacterial strains with increased nitrogen fixation rates able to secrete surplus nitrogen outside the cell, a crucial step in engineering improved nitrogen delivery to crop plants.

59 BASIC BIOLOGICAL SCIENCES↗

HAP1 gene expression is associated with radiosensitivity in breast cancer cells

Highlights: • Overexpression of HAP1 gene promotes apoptosis in MCF-7 cells after irradiation. • HAP1 reduces tumor volume in nude mice xenograft models after irradiation. • HAP1 increases radiosensitivity of breast cancer cells in vitro and vivo. - Abstract: Objectives: The purpose of this study was to investigate the relationship between huntingtin-associated protein1 (HAP1) gene and radiation therapy of breast cancer cells. Methods: HAP1 gene was transfected into breast cancer MCF-7 cells, which was confirmed by quantitative reverse transcription-polymerase chain reaction analysis (qRT-PCR) and Western blot in vitro. The changes of cell radiosensitivity were assessed by colony formation assay. Apoptosis were examined by flow cytometry. The expressions of two radiation-induced genes were evaluated by Western blot. Tumor growth was investigated in nude mice xenograft models in vivo. Results: Our data showed that HAP1 gene expression was significantly increased in HAP1-transfected MCF-7 cells in comparison with the parental cells or negative control cells. The survival rate in MCF-7/HAP1 cells was significantly decreased after irradiation (0, 2, 4, 6, 8 Gy), compared to cells in MCF-7 and MCF-7/Pb groups in vitro. HAP1 gene increased apoptosis in MCF-7 cells after irradiation. Additionally, the tumor volume and weight in MCF-7/HAP1 + RT group were observably lower than in MCF-7/HAP1 group and MCF-7/Pb + RT group. Conclusion: The present study indicated that HAP1 gene expression was related to the radiosensitivity of breast cancer cells and may play an important role in the regulation of cellular radiosensitivity.

60 APPLIED LIFE SCIENCES↗

Molecular insights into how a deficiency of amylose affects carbon allocation -- carbohydrate and oil analyses and gene expression profiling in the seeds of a rice waxy mutant

Background: Understanding carbon partitioning in cereal seeds is of critical importance to develop cereal crops with enhanced starch yields for food security and for producing specified end-products high in amylose, β-glucan, or fructan, such as functional foods or oils for biofuel applications. Waxy mutants of cereals have a high content of amylopectin and have been well characterized. However, the allocation of carbon to other components, such as β-glucan and oils, and the regulation of the altered carbon distribution to amylopectin in a waxy mutant are poorly understood. In this study, we used a rice mutant, GM077, with a low content of amylose to gain molecular insight into how a deficiency of amylose affects carbon allocation to other end products and to amylopectin. We used carbohydrate analysis, subtractive cDNA libraries, and qPCR to identify candidate genes potentially responsible for the changes in carbon allocation in GM077 seeds. Results: Carbohydrate analysis indicated that the content of amylose in GM077 seeds was significantly reduced, while that of amylopectin significantly rose as compared to the wild type BP034. The content of glucose, sucrose, total starch, cell-wall polysaccharides and oil were only slightly affected in the mutant as compared to the wild type. Suppression subtractive hybridization (SSH) experiments generated 116 unigenes in the mutant on the wild-type background. Among the 116 unigenes, three, AGP, ISA1 and SUSIBA2-like, were found to be directly involved in amylopectin synthesis, indicating their possible roles in redirecting carbon flux from amylose to amylopectin. A bioinformatics analysis of the putative SUSIBA2-like binding elements in the promoter regions of the upregulated genes indicated that the SUSIBA2-like transcription factor may be instrumental in promoting the carbon reallocation from amylose to amylopectin. Conclusion: Analyses of carbohydrate and oil fractions and gene expression profiling on a global scale in the rice waxy mutant GM077 revealed several candidate genes implicated in the carbon reallocation response to an amylose deficiency, including genes encoding AGPase and SUSIBA2-like. We believe that AGP and SUSIBA2 are two promising targets for classical breeding and/or transgenic plant improvement to control the carbon flux between starch and other components in cereal seeds.

54 ENVIRONMENTAL SCIENCES↗

Nuclear orphan receptor TLX affects gene expression, proliferation and cell apoptosis in beta cells

Nuclear orphan receptor TLX is an essential regulator of the growth of neural stem cells. However, its exact function in pancreatic islet cells is still unknown. In the present study, gene expression profiling analysis revealed that overexpression of TLX in beta cell line MIN6 causes suppression of 176 genes and upregulation of 49 genes, including a cadre of cell cycle, cell proliferation and cell death control genes, such as Btg2, Ddit3 and Gadd45a. We next examined the effects of TLX overexpression on proliferation, apoptosis and insulin secretion in MIN6 cells. Proliferation analysis using EdU assay showed that overexpression of TLX increased percentage of EdU-positive cells. Cell cycle and apoptosis analysis revealed that overexpression of TLX in MIN6 cells resulted in higher percentage of cells exiting G1 into S-phase, and a 58.8% decrease of cell apoptosis induced by 0.5 mM palmitate. Moreover, TLX overexpression did not cause impairment of insulin secretion. Together, we conclude that TLX is among factors capable of controlling beta cell proliferation and survival, which may serve as a target for the development of novel therapies for diabetes. - Highlights: • TLX overexpression in MIN6 cell causes significant expression changes of 225 genes. • TLX overexpression promotes MIN6 cell proliferation and decreases cell apoptosis. • TLX overexpression does not cause impairment of insulin secretion.

60 APPLIED LIFE SCIENCES↗

Muscle fiber type specific induction of slow myosin heavy chain 2 gene expression by electrical stimulation

Vertebrate skeletal muscle fiber types are defined by a broad array of differentially expressed contractile and metabolic protein genes. The mechanisms that establish and maintain these different fiber types vary throughout development and with changing functional demand. Chicken skeletal muscle fibers can be generally categorized as fast and fast/slow based on expression of the slow myosin heavy chain 2 (MyHC2) gene in fast/slow muscle fibers. To investigate the cellular and molecular mechanisms that control fiber type formation in secondary or fetal muscle fibers, myoblasts from the fast pectoralis major (PM) and fast/slow medial adductor (MA) muscles were isolated, allowed to differentiate in vitro, and electrically stimulated. MA muscle fibers were induced to express the slow MyHC2 gene by electrical stimulation, whereas PM muscle fibers did not express the slow MyHC2 gene under identical stimulation conditions. However, PM muscle fibers did express the slow MyHC2 gene when electrical stimulation was combined with inhibition of inositol triphosphate receptor (IP3R) activity. Electrical stimulation was sufficient to increase nuclear localization of expressed nuclear-factor-of-activated-T-cells (NFAT), NFAT-mediated transcription, and slow MyHC2 promoter activity in MA muscle fibers. In contrast, both electrical stimulation and inhibitors of IP3R activity were required for these effects in PM muscle fibers. Electrical stimulation also increased levels of peroxisome-proliferator-activated receptor-{gamma} co-activator-1 (PGC-1{alpha}) protein in PM and MA muscle fibers. These results indicate that MA muscle fibers can be induced by electrical stimulation to express the slow MyHC2 gene and that fast PM muscle fibers are refractory to stimulation-induced slow MyHC2 gene expression due to fast PM muscle fiber specific cellular mechanisms involving IP3R activity.

60 APPLIED LIFE SCIENCES↗

Age-related gene expression in luminal epithelial cells is driven by a microenvironment made from myoepithelial cells

Luminal epithelial cells in the breast gradually alter gene and protein expression with age, appearing to lose lineage-specificity by acquiring myoepithelial-like characteristics. We hypothesize that the luminal lineage is particularly sensitive to microenvironment changes, and age-related microenvironment changes cause altered luminal cell phenotypes. To evaluate the effects of different microenvironments on the fidelity of epigenetically regulated luminal and myoepithelial gene expression, we generated a set of lineage-specific probes for genes that are controlled through DNA methylation. Culturing primary luminal cells under conditions that favor myoepithelial propogation led to their reprogramming at the level of gene methylation, and to a more myoepithelial-like expression profile. Primary luminal cells’ lineage-specific gene expression could be maintained when they were cultured as bilayers with primary myoepithelial cells. Isogenic stromal fibroblast co-cultures were unable to maintain the luminal phenotype. Mixed-age luminal-myoepithelial bilayers revealed that luminal cells adopt transcription and methylation patterns consistent with the chronological age of the myoepithelial cells. We provide evidence that the luminal epithelial phenotype is exquisitely sensitive to microenvironment conditions, and that states of aging are cell non-autonomously communicated through microenvironment cues over at least one cell diameter.

59 BASIC BIOLOGICAL SCIENCES↗

Conducting On-orbit Gene Expression Analysis on ISS: WetLab-2

WetLab-2 will enable expanded genomic research on orbit by developing tools that support in situ sample collection, processing, and analysis on ISS. This capability will reduce the time-to-results for investigators and define new pathways for discovery on the ISS National Lab. The primary objective is to develop a research platform on ISS that will facilitate real-time quantitative gene expression analysis of biological samples collected on orbit. WetLab-2 will be capable of processing multiple sample types ranging from microbial cultures to animal tissues dissected on orbit. WetLab-2 will significantly expand the analytical capabilities onboard ISS and enhance science return from ISS.

quantitative PCR↗

Chitosan oligosaccharides enhance lipid droplets via down-regulation of PCSK9 gene expression in HepG2 cells

Highlights: • COS enhance lipid droplets in HepG2 cells. • COS down-regulate the expression of PCSK9 to maintain hepatic LDLR abundance and activity. • COS inhibit HNF1α and SREBP-2 activity leading to down-regulation of PCSK9. Chitosan oligosaccharides (COS), linear polymers of N-acetyl-D-glucosamine and deacetylated glucosamine, exhibit diverse pharmacological effects such as antimicrobial, antitumor, antioxidant and anti-inflammatory activities. Here, we explored their hypocholesterolemic effects in vivo and the molecular mechanisms of COS in hepatic cells. Our in vivo study of dyslipidemic ApoE{sup -/-} male mice showed that COS treatment of 500 mg kg{sup −1} d{sup −1} for 4 weeks clearly reduced the lipid deposits in the aorta and significantly decreased the hepatic proprotein convertase subtilisin/kexin type 9 (PCSK9) protein levels versus HFD groups (p < 0.05). To elucidate the mechanisms behind these effects, the HepG2 cell line was treated with COS. We found that COS (200 μg/ml) increased the amount of cell-surface low-density lipoprotein receptor (LDLR) and enhanced the lipid droplets in HepG2 cells (p < 0.05). The mRNA levels of LDLR and HMG-CoA protein levels were not altered, and the mRNA levels of PCSK9 were down-regulated by COS treatment for 24 h. We also observed that the expression levels of SREBP-2 (125 kD) and HNF-1α were increased in total cell lysates, but nuclear SREBP-2 (nSREBP-2, 68 kD, the active subunit of SREBP-2) levels were decreased and FOXO3a levels increased in nuclear lysates after COS treatment for 24 h. We demonstrated that one of the reasons for regulation of lipid transfer with COS is that FOXO3a levels are up-regulated by COS, leading to a reduction in the PCSK9 promoter binding capacity of HNF-1α and thus suppressing PCSK9 gene expression, up-regulating LDLR levels, and enhancing the lipid droplets in HepG2 cells. In addition, decreased expression of the PCSK9 gene was also contributed to by down-regulation of SREBP-2 by COS. We further confirmed the effect of suppression of PCSK9 expression by COS by utilizing RNA interference to silence HNF-1α and SREBP-2. Finally, to the best of our knowledge, we are the first to demonstrate that PCSK9 expression and LDLR activity are synergistically changed by a combination of HNF-1α and SREBP-2 after COS treatment. Our findings indicate that COS may regulate PCSK9 to modulate hepatic LDLR abundance and activity.

60 APPLIED LIFE SCIENCES↗

Cell-Type-Specific Profiling of Gene Expression and Chromatin Binding without Cell Isolation: Assaying RNA Pol II Occupancy in Neural Stem Cells

Cell-type-specific transcriptional profiling often requires the isolation of specific cell types from complex tissues. We have developed ‘‘TaDa,’’ a technique that enables cell-specific profiling without cell isolation. TaDa permits genome-wide profiling of DNA- or chromatin-binding proteins without cell sorting, fixation, or affinity purification. The method is simple, sensitive, highly reproducible, and transferable to any model system. We show that TaDa can be used to identify transcribed genes in a celltype-specific manner with considerable temporal precision, enabling the identification of differential gene expression between neuroblasts and the neuroepithelial cells from which they derive. We profile the genome-wide binding of RNA polymerase II in these adjacent, clonally related stem cells within intact Drosophila brains. Our data reveal expression of specific metabolic genes in neuroepithelial cells, but not in neuroblasts, and highlight gene regulatory networks that may pattern neural stem cell fates.

59 BASIC BIOLOGICAL SCIENCES↗

Data augmentation and multimodal learning for predicting drug response in patient-derived xenografts from gene expressions and histology images

Patient-derived xenografts (PDXs) are an appealing platform for preclinical drug studies. A primary challenge in modeling drug response prediction (DRP) with PDXs and neural networks (NNs) is the limited number of drug response samples. We investigate multimodal neural network (MM-Net) and data augmentation for DRP in PDXs. The MM-Net learns to predict response using drug descriptors, gene expressions (GE), and histology whole-slide images (WSIs). We explore whether combining WSIs with GE improves predictions as compared with models that use GE alone. We propose two data augmentation methods which allow us training multimodal and unimodal NNs without changing architectures with a single larger dataset: 1) combine single-drug and drug-pair treatments by homogenizing drug representations, and 2) augment drug-pairs which doubles the sample size of all drug-pair samples. Unimodal NNs which use GE are compared to assess the contribution of data augmentation. The NN that uses the original and the augmented drug-pair treatments as well as single-drug treatments outperforms NNs that ignore either the augmented drug-pairs or the single-drug treatments. In assessing the multimodal learning based on the MCC metric, MM-Net outperforms all the baselines. Our results show that data augmentation and integration of histology images with GE can improve prediction performance of drug response in PDXs.

60 APPLIED LIFE SCIENCES↗