A variational autoencoder trained with priors from canonical pathways increases the interpretability of transcriptome data.

Interpreting transcriptome data is an important yet challenging aspect of bioinformatic analysis. While gene set enrichment analysis is a standard tool for interpreting regulatory changes, we utilize deep learning techniques, specifically autoencoder architectures, to learn latent variables that drive transcriptome signals. We investigate whether simple, variational autoencoder (VAE), and beta-weighted VAE are capable of learning reduced representations of transcriptomes that retain critical biological information. We propose a novel VAE that utilizes priors from biological data to direct the network to learn a representation of the transcriptome that is based on understandable biological concepts. After benchmarking five different autoencoder architectures, we found that each succeeded in reducing the transcriptomes to 50 latent dimensions, which captured enough variation for accurate reconstruction. The simple, fully connected autoencoder, performs best across the benchmarks, but lacks the characteristic of having directly interpretable latent dimensions. The beta-weighted, prior-informed VAE implementation is able to solve the benchmarking tasks, and provide semantically accurate latent features equating to biological pathways. This study opens a new direction for differential pathway analysis in transcriptomics with increased transparency and interpretability.

Pleural CD14+ monocytes/macrophages of healthy adolescents show a high expression of metallothionein family genes.

Nowadays laparoscopic interventions enable the collection of resident macrophage populations out of the human cavities. We employed this technique to isolate pleural monocytes/macrophages from healthy young adults who underwent a correction of pectus excavatum. High quality CD14+ monocytes/macrophages (plMo/Mφ) were used for RNA-sequencing (RNA-seq) in comparison with human monocyte-derived macrophages (MDM) natural (MDM-0) or IL-4-polarized (MDM-IL4). Transcriptome analysis revealed 7166 and 7076 differentially expressed genes (DEGs) in plMo/Mφ relative to natural MDM-0 and polarized MDM-IL4, respectively. The gene set enrichment analysis, which was used to compare RNA-seq data from plMo/Mφ with single-cell (scRNA-seq) data online from human bronchial lavage macrophages, showed that plMo/Mφs are characterized by a high expression of genes belonging to the metallothionein (MT) family, and that the expression of these genes is significantly higher in plMo/Mφ than in MDM-0 or MDM-IL4. Our results provide additional insights on high MTs-expressing macrophage subsets, which seem to be present not only in bronchial lavage of healthy adults or in pleural exudates of lung cancer patients but also in pleural fluid of healthy young adults. Macrophage subsets expressing high MTs may have specific roles in lung defense, repair, and homeostasis, and require further investigations.

B Cells Are Not Involved in the Regulation of Adenoviral TGF-ß1- or Bleomycin-Induced Lung Fibrosis in Mice.

Idiopathic pulmonary fibrosis (IPF) is an irreversible, age-related diffuse parenchymal lung disease of poorly defined etiology. Many patients with IPF demonstrate distinctive lymphocytic interstitial infiltrations within remodeled lung tissue with uncertain pathogenetic relevance. Histopathological examination of explant lung tissue of patients with IPF revealed accentuated lymphoplasmacellular accumulations in close vicinity to, or even infiltrating, remodeled lung tissue. Similarly, we found significant accumulations of B cells interfused with T cells within remodeled lung tissue in two murine models of adenoviral TGF-ß1 or bleomycin (BLM)-induced lung fibrosis. Such B cell accumulations coincided with significantly increased lung collagen deposition, lung histopathology, and worsened lung function in wild-type (WT) mice. Surprisingly, B cell-deficient µMT knockout mice exhibited similar lung tissue remodeling and worsened lung function upon either AdTGF-ß1 or BLM as for WT mice. Comparative transcriptomic profiling of sorted B cells collected from lungs of AdTGF-ß1- and BLM-exposed WT mice identified a large set of commonly regulated genes, but with significant enrichment observed for Gene Ontology terms apparently not related to lung fibrogenesis. Collectively, although we observed B cell accumulations in lungs of IPF patients as well as two experimental models of lung fibrosis, comparative profiling of characteristic features of lung fibrosis between WT and B cell-deficient mice did not support a major involvement of B cells in lung fibrogenesis in mice.

Human Bronchial Epithelial Cell Transcriptome Changes in Response to Serum from Patients with Different Status of Inflammation.

PURPOSE: To investigate the transcriptome of human bronchial epithelial cells (HBEC) in response to serum from patients with different degrees of inflammation. METHODS: Serum from 19 COVID-19 patients obtained from the Hannover Unified Biobank was used. At the time of sampling, 5 patients had a WHO Clinical Progression Scale (WHO-CPS) score of 9 (severe illness). The remaining 14 patients had a WHO-CPS of below 9 (range 1-7), and lower illness. Multiplex immunoassay was used to assess serum inflammatory markers. The culture medium of HBEC was supplemented with 2% of the patient's serum, and the cells were cultured at 37 °C, 5% CO2 for 18 h. Subsequently, cellular RNA was used for RNA-Seq. RESULTS: Patients with scores below 9 had significantly lower albumin and serum levels of E-selectin, IL-8, and MCP-1 than patients with scores of 9. Principal component analysis based on 500 "core genes" of RNA-seq segregated cells into two subsets: exposed to serum from 4 (I) and 15 (II) patients. Cells from a subset (I) treated with serum from 4 patients with a score of 9 showed 5566 differentially expressed genes of which 2793 were up- and 2773 downregulated in comparison with cells of subset II treated with serum from 14 patients with scores between 1 and 7 and one with score = 9. In subset I cells, a higher expression of TLR4 and CXCL8 but a lower CDH1, ACE2, and HMOX1, and greater effects on genes involved in metabolic regulation, cytoskeletal organization, and kinase activity pathways were observed. CONCLUSION: This simple model could be useful to characterize patient serum and epithelial cell properties.

Single cell versus single nucleus: transcriptome differences in the murine kidney after ischemia-reperfusion injury.

The kidney is a complex organ, which consists of multiple components with highly diverse cell types. A detailed understanding of these cell types in health and disease is crucial for the future development of preventive and curative treatment strategies. In recent years, single-cell RNA sequencing (scRNAseq) and single-nucleus RNA sequencing (snRNAseq) technology has opened up completely new possibilities in investigating the variety of renal cell populations in physiological and pathological states. Here, we systematically assessed differences between scRNAseq and snRNAseq approaches in transcriptome analysis of murine kidneys after ischemia-reperfusion injury. We included tissues from control kidneys and from kidneys harvested 1 wk after mild (17-min clamping time) and severe (27-min clamping time) transient unilateral ischemia. Our findings revealed important methodological differences in the discovery of inflammatory cells, tubular cells, and other specialized cell types. Although the scRNAseq approach was advantageous for investigating immune cells, the snRNAseq approach allowed superior insights into healthy and damaged tubular cells. Apart from differences in the quantitative discovery rate, we found important qualitative discrepancies in the captured transcriptomes with crucial consequences for the interpretation of cell states and molecular functions. Together, we provide an overview of method-dependent differences between scRNAseq and snRNAseq results from identical postischemic kidney tissues. Our results highlight the importance of choosing the right approach for specific research questions.NEW & NOTEWORTHY Single-cell and single-nucleus RNA sequencing technologies provide powerful new tools to examine complex tissues such as the kidney. This research reference paper provides practical information on the differences between the two technologies when examining murine kidneys after ischemia-reperfusion injury. The results will serve those who are debating which protocols to use in their given study.

Loss of Serpina1 in Mice Leads to Altered Gene Expression in Inflammatory and Metabolic Pathways.

The SERPINA1 gene encodes alpha1-antitrypsin (AAT), an acute phase glycoprotein and serine protease inhibitor that is mainly (80-90%) produced in the liver. Point mutations in the SERPINA1 gene can lead to the misfolding, intracellular accumulation, and deficiency of circulating AAT protein, increasing the risk of developing chronic liver diseases or chronic obstructive pulmonary disease. Currently, siRNA technology can knock down the SERPINA1 gene and limit defective AAT production. How this latter affects other liver genes is unknown. Livers were taken from age- and sex-matched C57BL/6 wild-type (WT) and Serpina1 knockout mice (KO) aged from 8 to 14 weeks, all lacking the five serpin A1a-e paralogues. Total RNA was isolated and RNA sequencing, and transcriptome analysis was performed. The knockout of the Serpina1 gene in mice changed inflammatory, lipid metabolism, and cholesterol metabolism-related gene expression in the liver. Independent single-cell sequencing data of WT mice verified the involvement of Serpina1 in cholesterol metabolism. Our results from mice livers suggested that designing therapeutic strategies for the knockout of the SERPINA1 gene in humans must account for potential perturbations of key metabolic pathways and consequent mitigation of side effects.

MicroRNA-449a Inhibits Triple Negative Breast Cancer by Disturbing DNA Repair and Chromatid Separation.

Chromosomal instability (CIN) can be a driver of tumorigenesis but is also a promising therapeutic target for cancer associated with poor prognosis such as triple negative breast cancer (TNBC). The treatment of TNBC cells with defects in DNA repair genes with poly(ADP-ribose) polymerase inhibitor (PARPi) massively increases CIN, resulting in apoptosis. Here, we identified a previously unknown role of microRNA-449a in CIN. The transfection of TNBC cell lines HCC38, HCC1937 and HCC1395 with microRNA-449a mimics led to induced apoptosis, reduced cell proliferation, and reduced expression of genes in homology directed repair (HDR) in microarray analyses. EME1 was identified as a new target gene by immunoprecipitation and luciferase assays. The reduced expression of EME1 led to an increased frequency of ultrafine bridges, 53BP1 foci, and micronuclei. The induced expression of microRNA-449a elevated CIN beyond tolerable levels and induced apoptosis in TNBC cell lines by two different mechanisms: (I) promoting chromatid mis-segregation by targeting endonuclease EME1 and (II) inhibiting HDR by downregulating key players of the HDR network such as E2F3, BIRC5, BRCA2 and RAD51. The ectopic expression of microRNA-449a enhanced the toxic effect of PARPi in cells with pathogenic germline BRCA1 variants. The newly identified role makes microRNA-449a an interesting therapeutic target for TNBC.

IL-17 regulates DC migration to the peribronchial LNs and allergen presentation in experimental allergic asthma.

IL-17 is associated with different phenotypes of asthma, however, it is not fully elucidated how it influences induction and maintenance of asthma and allergy. In order to determine the role of IL-17 in development of allergic asthma, we used IL-17A/F double KO (IL-17A/F KO) and WT mice with or without neutralization of IL-17 in an experimental allergic asthma model and analyzed airway hyperresponsiveness, lung inflammation, T helper cell polarization, and DCs influx and activation. We report that the absence of IL-17 reduced influx of DCs into lungs and lung draining LNs. Compared to WT mice, IL-17A/F KO mice or WT mice after neutralization of IL-17A showed reduced airway hyperresponsiveness, eosinophilia, mucus hypersecretion, and IgE levels. DCs from draining LNs of allergen-challenged IL-17A/F KO mice showed a reduction in expression of migratory and costimulatory molecules CCR7, CCR2, MHC-II, and CD40 compared to WT DCs. Moreover, in vivo stimulation of adoptively transferred antigen-specific cells was attenuated in lung-draining LNs in the absence of IL-17. Thus, we report that IL-17 enhances airway DC activation, migration, and function. Consequently, lack of IL-17 leads to reduced antigen-specific T cell priming and impaired development of experimental allergic asthma.

A comparison of curated gene sets versus transcriptomics-derived gene signatures for detecting pathway activation in immune cells.

BACKGROUND: Despite the significant contribution of transcriptomics to the fields of biological and biomedical research, interpreting long lists of significantly differentially expressed genes remains a challenging step in the analysis process. Gene set enrichment analysis is a standard approach for summarizing differentially expressed genes into pathways or other gene groupings. Here, we explore an alternative approach to utilizing gene sets from curated databases. We examine the method of deriving custom gene sets which may be relevant to a given experiment using reference data sets from previous transcriptomics studies. We call these data-derived gene sets, "gene signatures" for the biological process tested in the previous study. We focus on the feasibility of this approach in analyzing immune-related processes, which are complicated in their nature but play an important role in the medical research. RESULTS: We evaluate several statistical approaches to detecting the activity of a gene signature in a target data set. We compare the performance of the data-derived gene signature approach with comparable GO term gene sets across all of the statistical tests. A total of 61 differential expression comparisons generated from 26 transcriptome experiments were included in the analysis. These experiments covered eight immunological processes in eight types of leukocytes. The data-derived signatures were used to detect the presence of immunological processes in the test data with modest accuracy (AUC = 0.67). The performance for GO and literature based gene sets was worse (AUC = 0.59). Both approaches were plagued by poor specificity. CONCLUSIONS: When investigators seek to test specific hypotheses, the data-derived signature approach can perform as well, if not better than standard gene-set based approaches for immunological signatures. Furthermore, the data-derived signatures can be generated in the cases that well-defined gene sets are lacking from pathway databases and also offer the opportunity for defining signatures in a cell-type specific manner. However, neither the data-derived signatures nor standard gene-sets can be demonstrated to reliably provide negative predictions for negative cases. We conclude that the data-derived signature approach is a useful and sometimes necessary tool, but analysts should be weary of false positives.

Complementary genomic approaches highlight the PI3K/mTOR pathway as a common vulnerability in osteosarcoma.

Osteosarcoma is the most common primary bone tumor, yet there have been no substantial advances in treatment or survival in three decades. We examined 59 tumor/normal pairs by whole-exome, whole-genome, and RNA-sequencing. Only the TP53 gene was mutated at significant frequency across all samples. The mean nonsilent somatic mutation rate was 1.2 mutations per megabase, and there was a median of 230 somatic rearrangements per tumor. Complex chains of rearrangements and localized hypermutation were detected in almost all cases. Given the intertumor heterogeneity, the extent of genomic instability, and the difficulty in acquiring a large sample size in a rare tumor, we used several methods to identify genomic events contributing to osteosarcoma survival. Pathway analysis, a heuristic analytic algorithm, a comparative oncology approach, and an shRNA screen converged on the phosphatidylinositol 3-kinase/mammalian target of rapamycin (PI3K/mTOR) pathway as a central vulnerability for therapeutic exploitation in osteosarcoma. Osteosarcoma cell lines are responsive to pharmacologic and genetic inhibition of the PI3K/mTOR pathway both in vitro and in vivo.

Comparative analysis of RNA sequencing methods for degraded or low-input samples.

RNA-seq is an effective method for studying the transcriptome, but it can be difficult to apply to scarce or degraded RNA from fixed clinical samples, rare cell populations or cadavers. Recent studies have proposed several methods for RNA-seq of low-quality and/or low-quantity samples, but the relative merits of these methods have not been systematically analyzed. Here we compare five such methods using metrics relevant to transcriptome annotation, transcript discovery and gene expression. Using a single human RNA sample, we constructed and sequenced ten libraries with these methods and compared them against two control libraries. We found that the RNase H method performed best for chemically fragmented, low-quality RNA, and we confirmed this through analysis of actual degraded samples. RNase H can even effectively replace oligo(dT)-based methods for standard RNA-seq. SMART and NuGEN had distinct strengths for measuring low-quantity RNA. Our analysis allows biologists to select the most suitable methods and provides a benchmark for future method development.

SF3B1 and other novel cancer genes in chronic lymphocytic leukemia.

BACKGROUND: The somatic genetic basis of chronic lymphocytic leukemia, a common and clinically heterogeneous leukemia occurring in adults, remains poorly understood. METHODS: We obtained DNA samples from leukemia cells in 91 patients with chronic lymphocytic leukemia and performed massively parallel sequencing of 88 whole exomes and whole genomes, together with sequencing of matched germline DNA, to characterize the spectrum of somatic mutations in this disease. RESULTS: Nine genes that are mutated at significant frequencies were identified, including four with established roles in chronic lymphocytic leukemia (TP53 in 15% of patients, ATM in 9%, MYD88 in 10%, and NOTCH1 in 4%) and five with unestablished roles (SF3B1, ZMYM3, MAPK1, FBXW7, and DDX3X). SF3B1, which functions at the catalytic core of the spliceosome, was the second most frequently mutated gene (with mutations occurring in 15% of patients). SF3B1 mutations occurred primarily in tumors with deletions in chromosome 11q, which are associated with a poor prognosis in patients with chronic lymphocytic leukemia. We further discovered that tumor samples with mutations in SF3B1 had alterations in pre-messenger RNA (mRNA) splicing. CONCLUSIONS: Our study defines the landscape of somatic mutations in chronic lymphocytic leukemia and highlights pre-mRNA splicing as a critical cellular process contributing to chronic lymphocytic leukemia.

An association between plasma levels of α2-macroglobulin and α1-antitrypsin in PiMM and PiZZ individuals differing in COPD presentation.

INTRODUCTION: Compared to normal PiMM, individuals with severe α1-antitrypsin (AAT) PiZZ (Glu342Lys) genotype deficiency are at higher risk of developing early-onset chronic obstructive pulmonary disease (COPD)/emphysema associated with Z-AAT polymers and neutrophilic inflammation. We aimed to investigate putative differences in plasma levels of acute phase proteins (APP) between PiMM and PiZZ subjects and to determine plasma Z-AAT polymer levels in PiZZ subjects. MATERIALS AND METHODS: Nephelometric analysis of seven plasma APPs was performed in 67 PiMM and 44 PiZZ subjects, of whom 43 and 42, respectively, had stable COPD. Of the PiZZ-COPD patients, 21 received and 23 did not receive intravenous therapy with human AAT preparations (IV-AAT). Plasma levels of Z-AAT polymers were determined by Western blotting using specific mouse monoclonal antibodies (2C1 and LG96). RESULTS: In addition to lower plasma AAT, PiZZ patients had higher α2-macroglobulin (A2MG) levels than PiMM patients. In contrast, PiZZ who received IV-AAT had higher AAT values but lower A2MG values than PiZZ without IV-AAT. Regardless of the AAT genotype, AAT levels were inversely correlated with A2MG, and the AAT/A2MG ratio was correlated with lung diffusion capacity (DCLO%). All PiZZ patients had circulating Z-AAT polymer levels that correlated directly with A2MG. In PiZZ without IV-AAT therapy polymer levels correlated inversely with the ratio of forced expiratory volume in 1 s to forced vital capacity (FEV1/FVC). CONCLUSION: Combined measurement of plasma AAT and A2MG levels may be of clinical value in assessing the progression of COPD and requires further attention.

Distinct metabolic responses to heme in inflammatory human and mouse macrophages - Role of nitric oxide.

Activation of inflammation is tightly associated with metabolic reprogramming in macrophages. The iron-containing tetrapyrrole heme can induce pro-oxidant and pro-inflammatory effects in murine macrophages, but has been associated with polarization towards an anti-inflammatory phenotype in human macrophages. In the current study, we compared the regulatory responses to heme and the prototypical Toll-like receptor (TLR)4 ligand lipopolysaccharide (LPS) in human and mouse macrophages with a particular focus on alterations of cellular bioenergetics. In human macrophages, bulk RNA-sequencing analysis indicated that heme led to an anti-inflammatory transcriptional profile, whereas LPS induced a classical pro-inflammatory gene response. Co-stimulation of heme with LPS caused opposing regulatory patterns of inflammatory activation and cellular bioenergetics in human and mouse macrophages. Specifically, in LPS-stimulated murine, but not human macrophages, heme led to a marked suppression of oxidative phosphorylation and an up-regulation of glycolysis. The species-specific alterations in cellular bioenergetics and inflammatory responses to heme were critically dependent on the availability of nitric oxide (NO) that is generated in inflammatory mouse, but not human macrophages. Accordingly, studies with an inducible nitric oxide synthase (iNOS) inhibitor in mouse, and a pharmacological NO donor in human macrophages, reveal that NO is responsible for the opposing effects of heme in these cells. Taken together, the current findings indicate that NO is critical for the immunomodulatory role of heme in macrophages.

RNA-SeQC: RNA-seq metrics for quality control and process optimization.

UNLABELLED: RNA-seq, the application of next-generation sequencing to RNA, provides transcriptome-wide characterization of cellular activity. Assessment of sequencing performance and library quality is critical to the interpretation of RNA-seq data, yet few tools exist to address this issue. We introduce RNA-SeQC, a program which provides key measures of data quality. These metrics include yield, alignment and duplication rates; GC bias, rRNA content, regions of alignment (exon, intron and intragenic), continuity of coverage, 3'/5' bias and count of detectable transcripts, among others. The software provides multi-sample evaluation of library construction protocols, input materials and other experimental parameters. The modularity of the software enables pipeline integration and the routine monitoring of key measures of data quality such as the number of alignable reads, duplication rates and rRNA contamination. RNA-SeQC allows investigators to make informed decisions about sample inclusion in downstream analysis. In summary, RNA-SeQC provides quality control measures critical to experiment design, process optimization and downstream computational analysis. AVAILABILITY AND IMPLEMENTATION: See www.genepattern.org to run online, or www.broadinstitute.org/rna-seqc/ for a command line tool.

Novel sequence feature variant type analysis of the HLA genetic association in systemic sclerosis.

We describe a novel approach to genetic association analyses with proteins sub-divided into biologically relevant smaller sequence features (SFs), and their variant types (VTs). SFVT analyses are particularly informative for study of highly polymorphic proteins such as the human leukocyte antigen (HLA), given the nature of its genetic variation: the high level of polymorphism, the pattern of amino acid variability, and that most HLA variation occurs at functionally important sites, as well as its known role in organ transplant rejection, autoimmune disease development and response to infection. Further, combinations of variable amino acid sites shared by several HLA alleles (shared epitopes) are most likely better descriptors of the actual causative genetic variants. In a cohort of systemic sclerosis patients/controls, SFVT analysis shows that a combination of SFs implicating specific amino acid residues in peptide binding pockets 4 and 7 of HLA-DRB1 explains much of the molecular determinant of risk.

Lung Adenocarcinoma Cell Sensitivity to Chemotherapies: A Spotlight on Lipid Droplets and SREBF1 Gene.

To explore the relationship between cancer cell SREBF1 expression, lipid droplets (LDs) formation, and the sensitivity to chemotherapies, we cultured lung adenocarcinoma cells H1299 (with LD) and H1563 (without LD) in a serum-free basal medium (BM) or neutrophil degranulation products containing medium (NDM), and tested cell responses to cisplatin and etoposide. By using the DESeq2 Bioconductor package, we detected 674 differentially expressed genes (DEGs) associated with NDM/BM differences between two cell lines, many of these genes were associated with the regulation of sterol and cholesterol biosynthesis processes. Specifically, SREBF1 markedly declined in both cell lines cultured in NDM or when treated with chemotherapeutics. Despite the latter, H1563 exhibited LD formation and resistance to etoposide, but not to cisplatin. Although H1299 cells preserved LDs, these cells were similarly sensitive to both drugs. In a cohort of 292 patients with non-small-cell lung cancer, a lower SREBF1 expression in tumors than in adjacent nontumor tissue correlated with overall better survival, specifically in patients with adenocarcinoma at stage I. Our findings imply that a direct correlation between SREBF1 and LD accumulation can be lost due to the changes in cancer cell environment and/or chemotherapy. The role of LDs in lung cancer development and response to therapies remains to be examined in more detail.

PB1-F2 Finder: scanning influenza sequences for PB1-F2 encoding RNA segments.

BACKGROUND: PB1-F2 is a major virulence factor of influenza A. This protein is a product of an alternative reading frame in the PB1-encoding RNA segment 2. Its presence of is dictated by the presence or absence of premature stop codons. This virulence factor is present in every influenza pandemic and major epidemic of the 20th century. Absence of PB1-F2 is associated with mild disease, such as the 2009 H1N1 ("swine flu"). RESULTS: The analysis of 8608 segment 2 sequences showed that only 8.5% have been annotated for the presence of PB1-F2. Our analysis indicates that 75% of segment 2 sequences are likely to encode PB1-F2. Two major populations of PB1-F2 are of lengths 90 and 57 while minor populations include lengths 52, 63, 79, 81, 87, and 101. Additional possible populations include the lengths of 59, 69, 81, 95, and 106. Previously described sequences include only lengths 57, 87, and 90. We observed substantial variation in PB1-F2 sequences where certain variants show up to 35% difference to well-defined reference sequences. Therefore this dataset indicates that there are many more variants that need to be functionally characterized. CONCLUSIONS: Our web-accessible tool PB1-F2 Finder enables scanning of influenza sequences for potential PB1-F2 protein products. It provides an initial screen and annotation of PB1-F2 products. It is accessible at http://cvc.dfci.harvard.edu/pb1-f2.

MaHCO: an ontology of the major histocompatibility complex for immunoinformatic applications and text mining.

MOTIVATION: The high level of polymorphism associated with the major histocompatibility complex (MHC) poses a challenge to organizing associated bioinformatic data, particularly in the area of hematopoietic stem cell transplantation. Thus, this area of research has great potential to profit from the ongoing development of biomedical ontologies, which offer structure and definition to MHC-data related communication and portability issues. RESULTS: We introduce the design considerations, methodological foundations and implementational issues underlying MaHCO, an ontology which represents the alleles and encoded molecules of the major histocompatibility complex. Importantly for human immunogenetics, it includes a detailed level of human leukocyte antigen (HLA) classification. We then present an ontology browser, search interfaces for immunogenetic fact and document retrieval, and the specification of an annotation language for semantic metadata, based on MaHCO. These use cases are intended to demonstrate the utility of ontology-driven bioinformatics in the field of immunogenetics. AVAILABILITY AND IMPLEMENTATION: The MaHCO Ontology is available via the BioPortal: http://www.bioontology.org/tools/portal/bioportal.html, and at: http://purl.org/stemnet/.

High-throughput minor histocompatibility antigen prediction.

MOTIVATION: Minor histocompatibility antigens (mHags) are a diverse collection of MHC-bound peptides that have immunological implications in the context of allogeneic transplantation because of their differential presence in donor and host, and thus play a critical role in the induction of the detrimental graft-versus-host disease (GvHD) or in the development of the beneficial graft-versus-leukemia (GvL) effect. Therefore, the search for mHags has implications not only for preventing GvHD, but also for therapeutic applications involving leukemia-specific T cells. We have created a web-based system, named PeptideCheck, which aims to augment the experimental discovery of mHags using bioinformatic means. Analyzing peptide elution data to search for mHags and predicting mHags from polymorphism and protein databases are the core features. RESULTS: Comparison with known mHag data reveals that some but not all of the previously known mHags can be reproduced. By applying a system of filtering and ranking, we were able to produce an ordered list of potential mHag candidates in which HA-1, HA-3 and HA-8 occur in the best 0.25%. By combining single nucleotide polymorphism, protein, tissue expression and genotypic frequency data, together with antigen presentation prediction algorithms, we propose a list of the best peptide candidates which could potentially induce the GvL effect without causing GvFD. AVAILABILITY: http://www.peptidecheck.org.