With the expansion of public repositories such as the Gene Expression Omnibus (GEO), we are rapidly cataloging cellular transcriptional responses to diverse experimental conditions. Methods that query these repositories based on gene expression content, rather than textual annotations, may enable more effective experiment retrieval as well as the discovery of novel associations between drugs, diseases, and other perturbations.

Serum proteins are routinely used to diagnose diseases, but are hard to find due to low sensitivity in screening the serum proteome. Public repositories of microarray data, such as the Gene Expression Omnibus (GEO), contain RNA expression profiles for more than 16,000 biological conditions, covering more than 30% of United States mortality. We hypothesized that genes coding for serum- and urine-detectable proteins, and showing differential expression of RNA in disease-damaged tissues would make ideal diagnostic protein biomarkers for those diseases. We showed that predicted protein biomarkers are significantly enriched for known diagnostic protein biomarkers in 22 diseases, with enrichment significantly higher in diseases for which at least three datasets are available. We then used this strategy to search for new biomarkers indicating acute rejection (AR) across different types of transplanted solid organs. We integrated three biopsy-based microarray studies of AR from pediatric renal, adult renal and adult cardiac transplantation and identified 45 genes upregulated in all three. From this set, we chose 10 proteins for serum ELISA assays in 39 renal transplant patients, and discovered three that were significantly higher in AR. Interestingly, all three proteins were also significantly higher during AR in the 63 cardiac transplant recipients studied. Our best marker, serum PECAM1, identified renal AR with 89% sensitivity and 75% specificity, and also showed increased expression in AR by immunohistochemistry in renal, hepatic and cardiac transplant biopsies. Our results demonstrate that integrating gene expression microarray measurements from disease samples and even publicly-available data sets can be a powerful, fast, and cost-effective strategy for the discovery of new diagnostic serum protein biomarkers.

With multiple genome-wide association studies (GWAS) performed across autoimmune diseases, there is a great opportunity to study the homogeneity of genetic architectures across autoimmune disease. Previous approaches have been limited in the scope of their analysis and have failed to properly incorporate the direction of allele-specific disease associations for SNPs. In this work, we refine the notion of a genetic variation profile for a given disease to capture strength of association with multiple SNPs in an allele-specific fashion. We apply this method to compare genetic variation profiles of six autoimmune diseases: multiple sclerosis (MS), ankylosing spondylitis (AS), autoimmune thyroid disease (ATD), rheumatoid arthritis (RA), Crohn's disease (CD), and type 1 diabetes (T1D), as well as five non-autoimmune diseases. We quantify pair-wise relationships between these diseases and find two broad clusters of autoimmune disease where SNPs that make an individual susceptible to one class of autoimmune disease also protect from diseases in the other autoimmune class. We find that RA and AS form one such class, and MS and ATD another. We identify specific SNPs and genes with opposite risk profiles for these two classes. We furthermore explore individual SNPs that play an important role in defining similarities and differences between disease pairs. We present a novel, systematic, cross-platform approach to identify allele-specific relationships between disease pairs based on genetic variation as well as the individual SNPs which drive the relationships. While recognizing similarities between diseases might lead to identifying novel treatment options, detecting differences between diseases previously thought to be similar may point to key novel disease-specific genes and pathways.

Cancer cell lines are used extensively to study cancer biology and to test hypotheses in translational research. The relevance of cell lines is dependent on how closely they resemble the tumors being studied. Relating tumors and cell lines, and recognizing their similarities and differences are thus very important for translational research. Rapid advances in genomics have led to the generation of large volumes of genomic and transcriptomic data for a diverse set of primary cancer samples, normal tissue samples and cancer cell lines. Hepatocellular Carcinoma (HCC) is one of the most common tumors worldwide, with high occurrence in Asia and sub-Saharan regions. The current effective treatments of HCC remain limited. In this work, we compared the gene expression measurements of 200 HCC tumor samples from The Cancer Genome Atlas and over 1000 cancer cell lines including 25 HCC cancer cell lines from Cancer Cell Line Encyclopedia. We showed that the HCC tumor samples correlate closely with HCC cell lines in comparison to cell lines derived from other tumor types. We further demonstrated that the most commonly used HCC cell lines resemble HCC tumors, while we identified nearly half of the cell lines that do not resemble primary tumors. Interestingly, a substantial number of genes that are critical for disease development or drug response are either expressed at low levels or absent among highly correlated cell lines; additional attention should be paid to these genes in translational research. Our study will be used to guide the selection of HCC cell lines and pinpoint the specific genes that are differentially expressed in either tumors or cell lines.

Type 2 diabetes (T2D) is a metabolic disease affecting >370 million people worldwide. It is characterized by obesity-induced insulin resistance, and growing evidence has indicated that this causative link between obesity and insulin resistance is associated with visceral adipose tissue inflammation. However, using anti-inflammatory drugs to treat insulin resistance and T2D is not a common practice. We recently applied a bioinformatics methodology to open public data and found that CD44 plays a critical role in the development of adipose tissue inflammation and insulin resistance. In this report, we examined the role of CD44 in T2D by administering daily injections of anti-CD44 monoclonal antibody (mAb) in a high-fat-diet mouse model. Four weeks of therapy with CD44 mAb suppressed visceral adipose tissue inflammation compared with controls and reduced fasting blood glucose levels, weight gain, liver steatosis, and insulin resistance to levels comparable to or better than therapy with the drugs metformin and pioglitazone. These findings suggest that CD44 mAb may be useful as a prototype drug for therapy of T2D by breaking the links between obesity and insulin resistance.

The daily timing of mammalian physiology is coordinated by circadian clocks throughout the body. Although measurements of clock gene expression indicate that these clocks in mice are normally in phase with each other, the situation in humans remains unclear. We used publicly available data from five studies, comprising over 1000 samples, to compare the phasing of circadian gene expression in human brain and human blood. Surprisingly, after controlling for age, clock gene expression in brain was phase-delayed by ~8.5 h relative to that of blood. We then examined clock gene expression in two additional human organs and in organs from nine other mammalian species, as well as in the suprachiasmatic nucleus (SCN). In most tissues outside the SCN, the expression of clock gene orthologs showed a phase difference of ~12 h between diurnal and nocturnal species. The exception to this pattern was human brain, whose phasing resembled that of the SCN. Our results highlight the value of a multi-tissue, multi-species meta-analysis, and have implications for our understanding of the human circadian system.

Complex diseases, such as Type 2 Diabetes Mellitus (T2D), result from the interplay of both environmental and genetic factors. However, most studies investigate either the genetics or the environment and there are a few that study their possible interaction in context of disease. One key challenge in documenting interactions between genes and environment includes choosing which of each to test jointly. Here, we attempt to address this challenge through a data-driven integration of epidemiological and toxicological studies. Specifically, we derive lists of candidate interacting genetic and environmental factors by integrating findings from genome-wide and environment-wide association studies. Next, we search for evidence of toxicological relationships between these genetic and environmental factors that may have an etiological role in the disease. We illustrate our method by selecting candidate interacting factors for T2D.

Long-term environmental variables are widely understood to play important roles in DNA variation. Previously, clinical studies examining the impacts of these variables on the human genome were localized to a single country, and used preselected DNA variants. Furthermore, clinical studies or surveys are either not available or difficult to carry out for developing countries. A systematic approach utilizing bioinformatics to identify associations among environmental variables, genetic variation, and diseases across various geographical locations is needed but has been lacking.

Human blood glucose levels have likely evolved toward their current point of stability over hundreds of thousands of years. The robust population stability of this trait is called canalization. It has been represented by a hyperbolic function of two variables: insulin sensitivity and insulin response. Environmental changes due to global migration may have pushed some human subpopulations to different points of stability. We hypothesized that there may be ethnic differences in the optimal states in the relationship between insulin sensitivity and insulin response.

Preeclampsia (PE) is a pregnancy-related vascular disorder which is the leading cause of maternal morbidity and mortality. We sought to identify novel serological protein markers to diagnose PE with a multi-'omics' based discovery approach.

Gene-expression profiling has become a mainstay in immunology, but subtle changes in gene networks related to biological processes are hard to discern when comparing various datasets. For instance, conservation of the transcriptional response to sepsis in mouse models and human disease remains controversial. To improve transcriptional analysis in immunology, we created ImmuneSigDB: a manually annotated compendium of ∼5,000 gene-sets from diverse cell states, experimental manipulations, and genetic perturbations in immunology. Analysis using ImmuneSigDB identified signatures induced in activated myeloid cells and differentiating lymphocytes that were highly conserved between humans and mice. Sepsis triggered conserved patterns of gene expression in humans and mouse models. However, we also identified species-specific biological processes in the sepsis transcriptional response: although both species upregulated phagocytosis-related genes, a mitosis signature was specific to humans. ImmuneSigDB enables granular analysis of transcriptomic data to improve biological understanding of immune processes of the human and mouse immune systems.

The central dogma of molecular biology describes the translation of genetic information from mRNA to protein, but does not specify the quantitation or timing of this process across the genome. We have analyzed protein and gene expression in a diverse set of human tissues. To study concordance and discordance of gene and protein expression, we integrated mass spectrometry data from the Human Proteome Map project and RNA-Seq measurements from the Genotype-Tissue Expression project. We analyzed 16,561 genes and the corresponding proteins in 14 tissue types across nearly 200 samples. A comprehensive tissue- and gene-specific analysis revealed that across the 14 tissues, correlation between mRNA and protein expression was positive and ranged from 0.36 to 0.5. We also identified 1,012 genes whose RNA and protein expression was correlated across all the tissues and examined genes and proteins that were concordantly and discordantly expressed for each tissue of interest. We extended our analysis to look for genes and proteins that were differentially correlated in cancer compared to normal tissues, showing higher levels of correlation in normal tissues. Finally, we explored the implications of these findings in the context of biomarker and drug target discovery.

Most GWASs were performed using study populations with Caucasian ethnicity or ancestry, and findings from one ethnic subpopulation might not always translate to another. We curated 4,573 genetic studies on 763 human diseases and identified 3,461 disease-susceptible SNPs with genome-wide significance; only 10% of these had been validated in at least two different ethnic populations. SNPs for autoimmune diseases demonstrated the lowest percentage of cross-ethnicity validation. We used the mortality data from the Center for Disease Control and Prevention and identified 19 diseases killing over 10,000 Americans per year that were still lacking publications of even a single cross-ethnic SNP. Fifteen of these diseases had never been studied in large GWAS in non-Caucasian populations, including chronic liver diseases and cirrhosis, leukemia, and non-Hodgkin's lymphoma. Our results demonstrate that diseases killing most Americans are still lacking genetic studies across ethnicities.

Preterm birth, or the delivery of an infant prior to 37 weeks of gestation, is a significant cause of infant morbidity and mortality. In the last decade, the advent and continued development of molecular profiling technologies has enabled researchers to generate vast amount of 'omics' data, which together with integrative computational approaches, can help refine the current knowledge about disease mechanisms, diagnostics, and therapeutics. Here we describe the March of Dimes' Database for Preterm Birth Research (http://www.immport.org/resources/mod), a unique resource that contains a variety of 'omics' datasets related to preterm birth. The database is open publicly, and as of January 2018, links 13 molecular studies with data across tens of thousands of patients from 6 measurement modalities. The data in the repository are highly diverse and include genomic, transcriptomic, immunological, and microbiome data. Relevant datasets are augmented with additional molecular characterizations of almost 25,000 biological samples from public databases. We believe our data-sharing efforts will lead to enhanced research collaborations and coordination accelerating the overall pace of discovery in preterm birth research.

The Gene Expression Omnibus (GEO) contains more than two million digital samples from functional genomics experiments amassed over almost two decades. However, individual sample meta-data remains poorly described by unstructured free text attributes preventing its largescale reanalysis. We introduce the Search Tag Analyze Resource for GEO as a web application (http://STARGEO.org) to curate better annotations of sample phenotypes uniformly across different studies, and to use these sample annotations to define robust genomic signatures of disease pathology by meta-analysis. In this paper, we target a small group of biomedical graduate students to show rapid crowd-curation of precise sample annotations across all phenotypes, and we demonstrate the biological validity of these crowd-curated annotations for breast cancer. STARGEO.org makes GEO data findable, accessible, interoperable and reusable (i.e., FAIR) to ultimately facilitate knowledge discovery. Our work demonstrates the utility of crowd-curation and interpretation of open 'big data' under FAIR principles as a first step towards realizing an ideal paradigm of precision medicine.

In order to advance precision medicine, detailed clinical features ought to be described in a way that leverages current knowledge. Although data collected from biomedical research is expanding at an almost exponential rate, our ability to transform that information into patient care has not kept at pace. A major barrier preventing this transformation is that multi-dimensional data collection and analysis is usually carried out without much understanding of the underlying knowledge structure. Here, in an effort to bridge this gap, Electronic Health Records (EHRs) of individual patients are connected to a heterogeneous knowledge network called Scalable Precision Medicine Oriented Knowledge Engine (SPOKE). Then an unsupervised machine-learning algorithm creates Propagated SPOKE Entry Vectors (PSEVs) that encode the importance of each SPOKE node for any code in the EHRs. We argue that these results, alongside the natural integration of PSEVs into any EHR machine-learning platform, provide a key step toward precision medicine.

Ovarian cancer is usually diagnosed at late stages when cancer has spread beyond the ovary and patients ultimately succumb to the development of drug-resistant disease. There is an urgent and unmet need to develop therapeutic strategies that effectively treat ovarian cancer and this requires a better understanding of signaling pathways important for ovarian cancer progression. Aurora A kinase (AURKA) plays an important role in ovarian cancer progression by mediating mitosis and chromosomal instability. In the current study, we investigated the role of AURKA in regulating the DNA damage response and DNA repair in ovarian carcinoma cells. We discovered that AURKA modulated the expression and activity of PARP, a crucial mediator of DNA repair that is a target of therapeutic interest for the treatment of ovarian and other cancers. Further, specific inhibition of AURKA activity with the small molecule inhibitor, alisertib, stimulated the non-homologous end-joining (NHEJ) repair pathway by elevating DNA-PKcs activity, a catalytic subunit required for double-strand break (DSB) repair, as well as decreased the expression of PARP and BRCA1/2, which are required for high-fidelity homologous recombination-based DNA repair. Further, AURKA inhibition stimulates error-prone NHEJ repair of DNA double-strand breaks with incompatible ends. Consistent with in vitro findings, alisertib treatment increased phosphorylated DNA-PKcs(pDNA-PKcsT2609) and decreased PARP levels in vivo. Collectively, these results reveal new non-mitotic functions for AURKA in the regulation of DNA repair, which may inform of new therapeutic targets and strategies for treating ovarian cancer.

Management of the COVID-19 pandemic has proven to be a significant challenge to policy makers. This is in large part due to uneven reporting and the absence of open-access visualization tools to present local trends and infer healthcare needs. Here we report the development of CovidCounties.org, an interactive web application that depicts daily disease trends at the level of US counties using time series plots and maps. This application is accompanied by a manually curated dataset that catalogs all major public policy actions made at the state-level, as well as technical validation of the primary data. Finally, the underlying code for the site is also provided as open source, enabling others to validate and learn from this work.

Tumor-infiltrating lymphocyte (TIL) therapy is a personalized cancer treatment which involves generating ex vivo cultures of tumor-reactive T cells from surgically resected tumors and administering the expanded TILs as a therapeutic infusion. Phase 1 of many TIL production protocols use aldesleukin (IL-2) alone to establish TIL cultures (termed "PreREP" (Pre-Rapid Expansion Protocol)); however, this fails to consistently produce TIL cultures from renal cell carcinoma (RCC) in a timely manner. Adding mitogenic stimulation via anti-CD3/anti-CD28 beads along with IL-2 to the fresh tumor digest (FTD) during TIL generation (termed "FTD+ beads") increases successful TIL culture rates; however, T cells produced by this method may be suboptimal for adoptive transfer. We hypothesize that adherent cell depletion (ACD) before TIL expansion will produce a superior TIL product by removing the immunosuppressive signals originating from adherent tumor and stromal cells. Here we investigate if "panning," a technique for ACD prior to TIL expansion, will impact the phenotype, functionality and/or clonality of ex vivo expanded RCC TILs.

The identification of novel disease associations using big-data for patient care has had limited success. In this study, we created a longitudinal disease network of traced readmissions (disease trajectories), merging data from over 10.4 million inpatients through the Healthcare Cost and Utilization Project, which allowed the representation of disease progression mapping over 300 diseases. From these disease trajectories, we discovered an interesting association between schizophrenia and rhabdomyolysis, a rare muscle disease (incidence < 1E-04) (relative risk, 2.21 [1.80-2.71, confidence interval = 0.95], P-value 9.54E-15). We validated this association by using independent electronic medical records from over 830,000 patients at the University of California, San Francisco (UCSF) medical center. A case review of 29 rhabdomyolysis incidents in schizophrenia patients at UCSF demonstrated that 62% are idiopathic, without the use of any drug known to lead to this adverse event, suggesting a warning to physicians to watch for this unexpected risk of schizophrenia. Large-scale analysis of disease trajectories can help physicians understand potential sequential events in their patients.