Precision Medicine: Improving health through high-resolution analysis of personal data.

For the 2019 Pacific Symposium on Biocomputing's session on precision medicine, we present new research on computational techniques in range of areas including data curation, whole genome analysis, transcriptomics, microbiome profiling, EHR data-mining, and histological image processing.

PRECISION MEDICINE: FROM DIPLOTYPES TO DISPARITIES TOWARDS IMPROVED HEALTH AND THERAPIES.

Precision medicine research efforts both in basic science discovery and clinical implementation are well underway and promise to provide individualized preventions and treatments, improving overall health care delivery. To achieve these goals, advances in data capture and analysis are needed spanning different types of 'omic and clinical data. The efforts to enhance precise treatments for all may accentuate healthcare disparities unless specific challenges are identified and addressed. This session of the 2018 Pacific Symposium on Biocomputing presents the latest developments in this transdisciplinary research space of genomics, medicine, and population health.

PRECISION MEDICINE: DATA AND DISCOVERY FOR IMPROVED HEALTH AND THERAPY.

The major goal of precision medicine is to improve human health. A feature that unites much research in the field is the use of large datasets such as genomic data and electronic health records. Research in this field includes examination of variation in the core bases of DNA and their methylation status, through variations in metabolic and signaling molecules, all the way up to broader systems level changes in physiology and disease presentation. Intermediate goals include understanding the individual drivers of disease that differentiate the cause of disease in each individual. To match this development of approaches to physical and activitybased measurements, computational approaches to using these new streams of data to better understand improve human health are being rapidly developed by the thriving biomedical informatics research community. This session of the 2017 Pacific Symposium of Biocomputing presents some of the latest advances in the capture, analysis and use of diverse biomedical data in precision medicine.

PRECISION MEDICINE: DATA AND DISCOVERY FOR IMPROVED HEALTH AND THERAPY.

Rapid advances in personal, cohort, and population-scale data acquisition, such as via sequencing, proteomics, mass spectroscopy, biosensors, mobile health devices and social network activity and other apps are opening up new vistas for personalized health biomedical data collection, analysis and insight. To achieve the vaunted goals of precision medicine and go from measurement to clinical translation, substantial gains still need to be made in methods of data and knowledge integration, analysis, discovery and interpretation. In this session of the 2016 Pacific Symposium on Biocomputing, we present sixteen papers to help accomplish this for precision medicine.

Disease Variant Landscape of a Large Multiethnic Population of Moyamoya Patients by Exome Sequencing.

Moyamoya disease (MMD) is a rare disorder characterized by cerebrovascular occlusion and development of hemorrhage-prone collateral vessels. Approximately 10-12% of cases are familial, with a presumed low penetrance autosomal dominant pattern of inheritance. Diagnosis commonly occurs only after clinical presentation. The recent identification of the RNF213 founder mutation (p.R4810K) in the Asian population has made a significant contribution, but the etiology of this disease remains unclear. To further develop the variant landscape of MMD, we performed high-depth whole exome sequencing of 125 unrelated, predominantly nonfamilial, ethnically diverse MMD patients in parallel with 125 internally sequenced, matched controls using the same exome and analysis platform. Three subpopulations were established: Asian, Caucasian, and non-RNF213 founder mutation cases. We provided additional support for the previously observed RNF213 founder mutation (p.R4810K) in Asian cases (P = 6.01×10(-5)) that was enriched among East Asians compared to Southeast Asian and Pacific Islander cases (P = 9.52×10(-4)) and was absent in all Caucasian cases. The most enriched variant in Caucasian (P = 7.93×10(-4)) and non-RNF213 founder mutation (P = 1.51×10(-3)) cases was ZXDC (p.P562L), a gene involved in MHC Class II activation. Collapsing variant methodology ranked OBSCN, a gene involved in myofibrillogenesis, as most enriched in Caucasian (P = 1.07×10(-4)) and non-RNF213 founder mutation cases (P = 5.31×10(-5)). These findings further support the East Asian origins of the RNF213 (p.R4810K) variant and more fully describe the genetic landscape of multiethnic MMD, revealing novel, alternative candidate variants and genes that may be important in MMD etiology and diagnosis.

Mouse models rarely mimic the transcriptome of human neurodegenerative diseases: A systematic bioinformatics-based critique of preclinical models.

Translational research for neurodegenerative disease depends intimately upon animal models. Unfortunately, promising therapies developed using mouse models mostly fail in clinical trials, highlighting uncertainty about how well mouse models mimic human neurodegenerative disease at the molecular level. We compared the transcriptional signature of neurodegeneration in mouse models of Alzheimer׳s disease (AD), Parkinson׳s disease (PD), Huntington׳s disease (HD) and amyotrophic lateral sclerosis (ALS) to human disease. In contrast to aging, which demonstrated a conserved transcriptome between humans and mice, only 3 of 19 animal models showed significant enrichment for gene sets comprising the most dysregulated up- and down-regulated human genes. Spearman׳s correlation analysis revealed even healthy human aging to be more closely related to human neurodegeneration than any mouse model of AD, PD, ALS or HD. Remarkably, mouse models frequently upregulated stress response genes that were consistently downregulated in human diseases. Among potential alternate models of neurodegeneration, mouse prion disease outperformed all other disease-specific models. Even among the best available animal models, conserved differences between mouse and human transcriptomes were found across multiple animal model versus human disease comparisons, surprisingly, even including aging. Relative to mouse models, mouse disease signatures demonstrated consistent trends toward preserved mitochondrial function protein catabolism, DNA repair responses, and chromatin maintenance. These findings suggest a more complex and multifactorial pathophysiology in human neurodegeneration than is captured through standard animal models, and suggest that even among conserved physiological processes such as aging, mice are less prone to exhibit neurodegeneration-like changes. This work may help explain the poor track record of mouse-based translational therapies for neurodegeneration and provides a path forward to critically evaluate and improve animal models of human disease.

Aging-like changes in the transcriptome of irradiated microglia.

Whole brain irradiation remains important in the management of brain tumors. Although necessary for improving survival outcomes, cranial irradiation also results in cognitive decline in long-term survivors. A chronic inflammatory state characterized by microglial activation has been implicated in radiation-induced brain injury. We here provide the first comprehensive transcriptional profile of irradiated microglia. Fluorescence-activated cell sorting was used to isolate CD11b+ microglia from the hippocampi of C57BL/6 and Balb/c mice 1 month after 10 Gy cranial irradiation. Affymetrix gene expression profiles were evaluated using linear modeling and rank product analyses. One month after irradiation, a conserved irradiation signature across strains was identified, comprising 448 and 85 differentially up- and downregulated genes, respectively. Gene set enrichment analysis demonstrated enrichment for inflammation, including M1 macrophage-associated genes, but also an unexpected enrichment for extracellular matrix and blood coagulation-related gene sets, in contrast previously described microglial states. Weighted gene coexpression network analysis confirmed these findings and further revealed alterations in mitochondrial function. The RNA-seq transcriptome of microglia 24-h postradiation proved similar to the 1-month transcriptome, but additionally featured alterations in apoptotic and lysosomal gene expression. Reanalysis of published aging mouse microglia transcriptome data demonstrated striking similarity to the 1-month irradiated microglia transcriptome, suggesting that shared mechanisms may underlie aging and chronic irradiation-induced cognitive decline. GLIA 2015;63:754-767.

Prediction of multiple infections after severe burn trauma: a prospective cohort study.

OBJECTIVE: To develop predictive models for early triage of burn patients based on hypersusceptibility to repeated infections. BACKGROUND: Infection remains a major cause of mortality and morbidity after severe trauma, demanding new strategies to combat infections. Models for infection prediction are lacking. METHODS: Secondary analysis of 459 burn patients (≥16 years old) with 20% or more total body surface area burns recruited from 6 US burn centers. We compared blood transcriptomes with a 180-hour cutoff on the injury-to-transcriptome interval of 47 patients (≤1 infection episode) to those of 66 hypersusceptible patients [multiple (≥2) infection episodes (MIE)]. We used LASSO regression to select biomarkers and multivariate logistic regression to built models, accuracy of which were assessed by area under receiver operating characteristic curve (AUROC) and cross-validation. RESULTS: Three predictive models were developed using covariates of (1) clinical characteristics; (2) expression profiles of 14 genomic probes; (3) combining (1) and (2). The genomic and clinical models were highly predictive of MIE status [AUROCGenomic = 0.946 (95% CI: 0.906-0.986); AUROCClinical = 0.864 (CI: 0.794-0.933); AUROCGenomic/AUROCClinical P = 0.044]. Combined model has an increased AUROCCombined of 0.967 (CI: 0.940-0.993) compared with the individual models (AUROCCombined/AUROCClinical P = 0.0069). Hypersusceptible patients show early alterations in immune-related signaling pathways, epigenetic modulation, and chromatin remodeling. CONCLUSIONS: Early triage of burn patients more susceptible to infections can be made using clinical characteristics and/or genomic signatures. Genomic signature suggests new insights into the pathophysiology of hypersusceptibility to infection may lead to novel potential therapeutic or prophylactic targets.

Integrated multi-cohort transcriptional meta-analysis of neurodegenerative diseases.

INTRODUCTION: Neurodegenerative diseases share common pathologic features including neuroinflammation, mitochondrial dysfunction and protein aggregation, suggesting common underlying mechanisms of neurodegeneration. We undertook a meta-analysis of public gene expression data for neurodegenerative diseases to identify a common transcriptional signature of neurodegeneration. RESULTS: Using 1,270 post-mortem central nervous system tissue samples from 13 patient cohorts covering four neurodegenerative diseases, we identified 243 differentially expressed genes, which were similarly dysregulated in 15 additional patient cohorts of 205 samples including seven neurodegenerative diseases. This gene signature correlated with histologic disease severity. Metallothioneins featured prominently among differentially expressed genes, and functional pathway analysis identified specific convergent themes of dysregulation. MetaCore network analyses revealed various novel candidate hub genes (e.g. STAU2). Genes associated with M1-polarized macrophages and reactive astrocytes were strongly enriched in the meta-analysis data. Evaluation of genes enriched in neurons revealed 70 down-regulated genes, over half not previously associated with neurodegeneration. Comparison with aging brain data (3 patient cohorts, 221 samples) revealed 53 of these to be unique to neurodegenerative disease, many of which are strong candidates to be important in neuropathogenesis (e.g. NDN, NAP1L2). ENCODE ChIP-seq analysis predicted common upstream transcriptional regulators not associated with normal aging (REST, RBBP5, SIN3A, SP2, YY1, ZNF143, IKZF1). Finally, we removed genes common to neurodegeneration from disease-specific gene signatures, revealing uniquely robust immune response and JAK-STAT signaling in amyotrophic lateral sclerosis. CONCLUSIONS: Our results implicate pervasive bioenergetic deficits, M1-type microglial activation and gliosis as unifying themes of neurodegeneration, and identify numerous novel genes associated with neurodegenerative processes.

Multiplex meta-analysis of medulloblastoma expression studies with external controls.

We propose and discuss a method for doing gene expression meta-analysis (multiple datasets) across multiplex measurement modalities measuring the expression of many genes simultaneously (e.g. microarrays and RNAseq) using external control samples and a method of heterogeneity detection to identify and filter on comparable gene expression measurements. We demonstrate this approach on publicly available gene expression datasets from samples of medulloblastoma and normal cerebellar tissue and identify some potential new targets in the treatment of medulloblastoma.

Integrating multiple 'omics' analyses identifies serological protein biomarkers for preeclampsia.

BACKGROUND: 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. METHODS: Seven previous placental expression studies were combined for a multiplex analysis, and in parallel, two-dimensional gel electrophoresis was performed to compare serum proteomes in PE and control subjects. The combined biomarker candidates were validated with available ELISA assays using gestational age-matched PE (n=32) and control (n=32) samples. With the validated biomarkers, a genetic algorithm was then used to construct and optimize biomarker panels in PE assessment. RESULTS: In addition to the previously identified biomarkers, the angiogenic and antiangiogenic factors (soluble fms-like tyrosine kinase (sFlt-1) and placental growth factor (PIGF)), we found 3 up-regulated and 6 down-regulated biomakers in PE sera. Two optimal biomarker panels were developed for early and late onset PE assessment, respectively. CONCLUSIONS: Both early and late onset PE diagnostic panels, constructed with our PE biomarkers, were superior over sFlt-1/PIGF ratio in PE discrimination. The functional significance of these PE biomarkers and their associated pathways were analyzed which may provide new insights into the pathogenesis of PE.

A common rejection module (CRM) for acute rejection across multiple organs identifies novel therapeutics for organ transplantation.

Using meta-analysis of eight independent transplant datasets (236 graft biopsy samples) from four organs, we identified a common rejection module (CRM) consisting of 11 genes that were significantly overexpressed in acute rejection (AR) across all transplanted organs. The CRM genes could diagnose AR with high specificity and sensitivity in three additional independent cohorts (794 samples). In another two independent cohorts (151 renal transplant biopsies), the CRM genes correlated with the extent of graft injury and predicted future injury to a graft using protocol biopsies. Inferred drug mechanisms from the literature suggested that two FDA-approved drugs (atorvastatin and dasatinib), approved for nontransplant indications, could regulate specific CRM genes and reduce the number of graft-infiltrating cells during AR. We treated mice with HLA-mismatched mouse cardiac transplant with atorvastatin and dasatinib and showed reduction of the CRM genes, significant reduction of graft-infiltrating cells, and extended graft survival. We further validated the beneficial effect of atorvastatin on graft survival by retrospective analysis of electronic medical records of a single-center cohort of 2,515 renal transplant patients followed for up to 22 yr. In conclusion, we identified a CRM in transplantation that provides new opportunities for diagnosis, drug repositioning, and rational drug design.

Differentiating the roles of STAT5B and STAT5A in human CD4+ T cells.

STAT5A and STAT5B are highly homologous proteins whose distinctive roles in human immunity remain unclear. However, STAT5A sufficiency cannot compensate for STAT5B defects, and human STAT5B deficiency, a rare autosomal recessive primary immunodeficiency, is characterized by chronic lung disease, growth failure and autoimmunity associated with regulatory T cell (Treg) reduction. We therefore hypothesized that STAT5A and STAT5B play unique roles in CD4(+) T cells. Upon knocking down STAT5A or STAT5B in human primary T cells, we found differentially regulated expression of FOXP3 and IL-2R in STAT5B knockdown T cells and down-regulated Bcl-X only in STAT5A knockdown T cells. Functional ex vivo studies in homozygous STAT5B-deficient patients showed reduced FOXP3 expression with impaired regulatory function of STAT5B-null Treg cells, also of increased memory phenotype. These results indicate that STAT5B and STAT5A act partly as non-redundant transcription factors and that STAT5B is more critical for Treg maintenance and function in humans.

Whole genome sequencing in support of wellness and health maintenance.

BACKGROUND: Whole genome sequencing is poised to revolutionize personalized medicine, providing the capacity to classify individuals into risk categories for a wide range of diseases. Here we begin to explore how whole genome sequencing (WGS) might be incorporated alongside traditional clinical evaluation as a part of preventive medicine. The present study illustrates novel approaches for integrating genotypic and clinical information for assessment of generalized health risks and to assist individuals in the promotion of wellness and maintenance of good health. METHODS: Whole genome sequences and longitudinal clinical profiles are described for eight middle-aged Caucasian participants (four men and four women) from the Center for Health Discovery and Well Being (CHDWB) at Emory University in Atlanta. We report multivariate genotypic risk assessments derived from common variants reported by genome-wide association studies (GWAS), as well as clinical measures in the domains of immune, metabolic, cardiovascular, musculoskeletal, respiratory, and mental health. RESULTS: Polygenic risk is assessed for each participant for over 100 diseases and reported relative to baseline population prevalence. Two approaches for combining clinical and genetic profiles for the purposes of health assessment are then presented. First we propose conditioning individual disease risk assessments on observed clinical status for type 2 diabetes, coronary artery disease, hypertriglyceridemia and hypertension, and obesity. An approximate 2:1 ratio of concordance between genetic prediction and observed sub-clinical disease is observed. Subsequently, we show how more holistic combination of genetic, clinical and family history data can be achieved by visualizing risk in eight sub-classes of disease. Having identified where their profiles are broadly concordant or discordant, an individual can focus on individual clinical results or genotypes as they develop personalized health action plans in consultation with a health partner or coach. CONCLUSION: The CHDWB will facilitate longitudinal evaluation of wellness-focused medical care based on comprehensive self-knowledge of medical risks.

Analysis of the genetic basis of disease in the context of worldwide human relationships and migration.

Genetic diversity across different human populations can enhance understanding of the genetic basis of disease. We calculated the genetic risk of 102 diseases in 1,043 unrelated individuals across 51 populations of the Human Genome Diversity Panel. We found that genetic risk for type 2 diabetes and pancreatic cancer decreased as humans migrated toward East Asia. In addition, biliary liver cirrhosis, alopecia areata, bladder cancer, inflammatory bowel disease, membranous nephropathy, systemic lupus erythematosus, systemic sclerosis, ulcerative colitis, and vitiligo have undergone genetic risk differentiation. This analysis represents a large-scale attempt to characterize genetic risk differentiation in the context of migration. We anticipate that our findings will enable detailed analysis pertaining to the driving forces behind genetic risk differentiation.

Proline: the distribution, frequency, positioning, and common functional roles of proline and polyproline sequences in the human proteome.

Proline is an anomalous amino acid. Its nitrogen atom is covalently locked within a ring, thus it is the only proteinogenic amino acid with a constrained phi angle. Sequences of three consecutive prolines can fold into polyproline helices, structures that join alpha helices and beta pleats as architectural motifs in protein configuration. Triproline helices are participants in protein-protein signaling interactions. Longer spans of repeat prolines also occur, containing as many as 27 consecutive proline residues. Little is known about the frequency, positioning, and functional significance of these proline sequences. Therefore we have undertaken a systematic bioinformatics study of proline residues in proteins. We analyzed the distribution and frequency of 687,434 proline residues among 18,666 human proteins, identifying single residues, dimers, trimers, and longer repeats. Proline accounts for 6.3% of the 10,882,808 protein amino acids. Of all proline residues, 4.4% are in trimers or longer spans. We detected patterns that influence function based on proline location, spacing, and concentration. We propose a classification based on proline-rich, polyproline-rich, and proline-poor status. Whereas singlet proline residues are often found in proteins that display recurring architectural patterns, trimers or longer proline sequences tend be associated with the absence of repetitive structural motifs. Spans of 6 or more are associated with DNA/RNA processing, actin, and developmental processes. We also suggest a role for proline in Kruppel-type zinc finger protein control of DNA expression, and in the nucleation and translocation of actin by the formin complex.

FoxO6 regulates memory consolidation and synaptic function.

The FoxO family of transcription factors is known to slow aging downstream from the insulin/IGF (insulin-like growth factor) signaling pathway. The most recently discovered FoxO isoform in mammals, FoxO6, is highly enriched in the adult hippocampus. However, the importance of FoxO factors in cognition is largely unknown. Here we generated mice lacking FoxO6 and found that these mice display normal learning but impaired memory consolidation in contextual fear conditioning and novel object recognition. Using stereotactic injection of viruses into the hippocampus of adult wild-type mice, we found that FoxO6 activity in the adult hippocampus is required for memory consolidation. Genome-wide approaches revealed that FoxO6 regulates a program of genes involved in synaptic function upon learning in the hippocampus. Consistently, FoxO6 deficiency results in decreased dendritic spine density in hippocampal neurons in vitro and in vivo. Thus, FoxO6 may promote memory consolidation by regulating a program coordinating neuronal connectivity in the hippocampus, which could have important implications for physiological and pathological age-dependent decline in memory.

Clinical utility of sequence-based genotype compared with that derivable from genotyping arrays.

OBJECTIVE: We investigated the common-disease relevant information obtained from sequencing compared with that reported from genotyping arrays. MATERIALS AND METHODS: Using 187 publicly available individual human genomes, we constructed genomic disease risk summaries based on 55 common diseases with reported gene-disease associations in the research literature using two different risk models, one based on the product of likelihood ratios and the other on the allelic variant with the maximum associated disease risk. We also constructed risk profiles based on the single nucleotide polymorphisms (SNPs) of these individuals that could be measured or imputed from two common genotyping array platforms. RESULTS: We show that the model risk predictions derived from sequencing differ substantially from those obtained from the SNPs measured on commercially available genotyping arrays for several different non-monogenic diseases, although high density genotyping arrays give identical results for many diseases. CONCLUSIONS: Our approach may be used to compare the ability of different platforms to probe known genetic risks disease by disease.

Expression-based genome-wide association study links the receptor CD44 in adipose tissue with type 2 diabetes.

Type 2 diabetes (T2D) is a complex, polygenic disease affecting nearly 300 million people worldwide. T2D is primarily characterized by insulin resistance, and growing evidence has indicated the causative link between adipose tissue inflammation and the development of insulin resistance. Genetic association studies have successfully revealed a number of important genes consistently associated with T2D to date. However, these robust T2D-associated genes do not fully elucidate the mechanisms underlying the development and progression of the disease. Here, we report an alternative approach, gene expression-based genome-wide association study (eGWAS): searching for genes repeatedly implicated in functional microarray experiments (often publicly available). We performed an eGWAS across 130 independent experiments (totally 1,175 T2D case-control microarrays) to find additional genes implicated in the molecular pathogenesis of T2D and identified the immune-cell receptor CD44 as our top candidate (P = 8.5 × 10(-20)). We found CD44 deficiency in a diabetic mouse model ameliorates insulin resistance and adipose tissue inflammation and also found that anti-CD44 antibody treatment decreases blood glucose levels and adipose tissue macrophage accumulation in a high-fat, diet-fed mouse model. Further, in humans, we observed CD44 is expressed in inflammatory cells in obese adipose tissue and discovered serum CD44 levels were positively correlated with insulin resistance and glycemic control. CD44 likely plays a causative role in the development of adipose tissue inflammation and insulin resistance in rodents and humans. Genes repeatedly implicated in publicly available experimental data may have unique functionally important roles in T2D and other complex diseases.

Multiplex meta-analysis of RNA expression to identify genes with variants associated with immune dysfunction.

OBJECTIVE: We demonstrate a genome-wide method for the integration of many studies of gene expression of phenotypically similar disease processes, a method of multiplex meta-analysis. We use immune dysfunction as an example disease process. DESIGN: We use a heterogeneous collection of datasets across human and mice samples from a range of tissues and different forms of immunodeficiency. We developed a method integrating Tibshirani's modified t-test (SAM) is used to interrogate differential expression within a study and Fisher's method for omnibus meta-analysis to identify differentially expressed genes across studies. The ability of this overall gene expression profile to prioritize disease associated genes is evaluated by comparing against the results of a recent genome wide association study for common variable immunodeficiency (CVID). RESULTS: Our approach is able to prioritize genes associated with immunodeficiency in general (area under the ROC curve = 0.713) and CVID in particular (area under the ROC curve = 0.643). CONCLUSIONS: This approach may be used to investigate a larger range of failures of the immune system. Our method may be extended to other disease processes, using RNA levels to prioritize genes likely to contain disease associated DNA variants.