Use of Multiple Machine Learning Approaches for Selecting Urothelial Cancer-Specific DNA Methylation Biomarkers in Urine.

Diagnosing urothelial cancer (UCa) via invasive cystoscopy is painful, specifically in men, and can cause infection and bleeding. Because the UCa risk is higher for male patients, urinary non-invasive UCa biomarkers are highly desired to stratify men for invasive cystoscopy. We previously identified multiple DNA methylation sites in urine samples that detect UCa with a high sensitivity and specificity in men. Here, we identified the most relevant markers by employing multiple statistical approaches and machine learning (random forest, boosted trees, LASSO) using a dataset of 251 male UCa patients and 111 controls. Three CpG sites located in ALOX5, TRPS1 and an intergenic region on chromosome 16 have been concordantly selected by all approaches, and their combination in a single decision matrix for clinical use was tested based on their respective thresholds of the individual CpGs. The combination of ALOX5 and TRPS1 yielded the best overall sensitivity (61%) at a pre-set specificity of 95%. This combination exceeded both the diagnostic performance of the most sensitive bioinformatic approach and that of the best single CpG. In summary, we showed that overlap analysis of multiple statistical approaches identifies the most reliable biomarkers for UCa in a male collective. The results may assist in stratifying men for cystoscopy.

Characterization of peptide-protein relationships in protein ambiguity groups via bipartite graphs.

In bottom-up proteomics, proteins are enzymatically digested into peptides before measurement with mass spectrometry. The relationship between proteins and their corresponding peptides can be represented by bipartite graphs. We conduct a comprehensive analysis of bipartite graphs using quantified peptides from measured data sets as well as theoretical peptides from an in silico digestion of the corresponding complete taxonomic protein sequence databases. The aim of this study is to characterize and structure the different types of graphs that occur and to compare them between data sets. We observed a large influence of the accepted minimum peptide length during in silico digestion. When changing from theoretical peptides to measured ones, the graph structures are subject to two opposite effects. On the one hand, the graphs based on measured peptides are on average smaller and less complex compared to graphs using theoretical peptides. On the other hand, the proportion of protein nodes without unique peptides, which are a complicated case for protein inference and quantification, is considerably larger for measured data. Additionally, the proportion of graphs containing at least one protein node without unique peptides rises when going from database to quantitative level. The fraction of shared peptides and proteins without unique peptides as well as the complexity and size of the graphs highly depends on the data set and organism. Large differences between the structures of bipartite peptide-protein graphs have been observed between database and quantitative level as well as between analyzed species. In the analyzed measured data sets, the proportion of protein nodes without unique peptides ranged from 6.4% to 55.0%. This highlights the need for novel methods that can quantify proteins without unique peptides. The knowledge about the structure of the bipartite peptide-protein graphs gained in this study will be useful for the development of such algorithms.

A Current Encyclopedia of Bioinformatics Tools, Data Formats and Resources for Mass Spectrometry Lipidomics.

Mass spectrometry is a widely used technology to identify and quantify biomolecules such as lipids, metabolites and proteins necessary for biomedical research. In this study, we catalogued freely available software tools, libraries, databases, repositories and resources that support lipidomics data analysis and determined the scope of currently used analytical technologies. Because of the tremendous importance of data interoperability, we assessed the support of standardized data formats in mass spectrometric (MS)-based lipidomics workflows. We included tools in our comparison that support targeted as well as untargeted analysis using direct infusion/shotgun (DI-MS), liquid chromatography-mass spectrometry, ion mobility or MS imaging approaches on MS1 and potentially higher MS levels. As a result, we determined that the Human Proteome Organization-Proteomics Standards Initiative standard data formats, mzML and mzTab-M, are already supported by a substantial number of recent software tools. We further discuss how mzTab-M can serve as a bridge between data acquisition and lipid bioinformatics tools for interpretation, capturing their output and transmitting rich annotated data for downstream processing. However, we identified several challenges of currently available tools and standards. Potential areas for improvement were: adaptation of common nomenclature and standardized reporting to enable high throughput lipidomics and improve its data handling. Finally, we suggest specific areas where tools and repositories need to improve to become FAIRer.

A proteomics sample metadata representation for multiomics integration and big data analysis.

The amount of public proteomics data is rapidly increasing but there is no standardized format to describe the sample metadata and their relationship with the dataset files in a way that fully supports their understanding or reanalysis. Here we propose to develop the transcriptomics data format MAGE-TAB into a standard representation for proteomics sample metadata. We implement MAGE-TAB-Proteomics in a crowdsourcing project to manually curate over 200 public datasets. We also describe tools and libraries to validate and submit sample metadata-related information to the PRIDE repository. We expect that these developments will improve the reproducibility and facilitate the reanalysis and integration of public proteomics datasets.

Advanced Fiber Type-Specific Protein Profiles Derived from Adult Murine Skeletal Muscle.

Skeletal muscle is a heterogeneous tissue consisting of blood vessels, connective tissue, and muscle fibers. The last are highly adaptive and can change their molecular composition depending on external and internal factors, such as exercise, age, and disease. Thus, examination of the skeletal muscles at the fiber type level is essential to detect potential alterations. Therefore, we established a protocol in which myosin heavy chain isoform immunolabeled muscle fibers were laser microdissected and separately investigated by mass spectrometry to develop advanced proteomic profiles of all murine skeletal muscle fiber types. All data are available via ProteomeXchange with the identifier PXD025359. Our in-depth mass spectrometric analysis revealed unique fiber type protein profiles, confirming fiber type-specific metabolic properties and revealing a more versatile function of type IIx fibers. Furthermore, we found that multiple myopathy-associated proteins were enriched in type I and IIa fibers. To further optimize the assignment of fiber types based on the protein profile, we developed a hypothesis-free machine-learning approach, identified a discriminative peptide panel, and confirmed our panel using a public data set.

Computertomography-Based Prediction of Complete Response Following Neoadjuvant Chemoradiotherapy of Locally Advanced Rectal Cancer.

Therapeutic strategies for patients with locally advanced rectal cancer (LARC) who are achieving a pathological complete response (pCR) after neoadjuvant radio-chemotherapy (neoCRT) are being increasingly investigated. Recent trials challenge the current standard therapy of total mesorectal excision (TME). For some patients, the treatment strategy of "watch-and-wait" seems a preferable procedure. The key factor in determining individual treatment strategies following neoCRT is the precise evaluation of the tumor response. Contrast-enhanced computer tomography (ceCT) has proven its ability to discriminate benign and malign lesions in multiple cancers. In this study, we retrospectively analyzed the ceCT based density of LARC in 30 patients, undergoing neoCRT followed by TME. We compared the tumors´ pre- and post-neoCRT density and correlated the results to the amount of residual vital tumor cells in the resected tissue. Overall, the density decreased after neoCRT, with the highest decrease in patients achieving pCR. Densitometry demonstrated a specificity of 88% and sensitivity of 68% in predicting pCR. Thus, we claim that ceCT based densitometry is a useful tool in identifying patients with LARC who may benefit from a "watch-and-wait" strategy and suggest further prospective studies.

Important Issues in Planning a Proteomics Experiment: Statistical Considerations of Quantitative Proteomic Data.

Mass spectrometry is frequently used in quantitative proteomics to detect differentially regulated proteins. A very important but unfortunately oftentimes neglected part in detecting differential proteins is the statistical analysis. Data from proteomics experiments are usually high-dimensional and hence require profound statistical methods. It is especially important to already correctly design a proteomic experiment before it is conducted in the laboratory. Only this can ensure that the statistical analysis is capable of detecting truly differential proteins afterward. This chapter thus covers aspects of both statistical planning as well as the actual analysis of quantitative proteomic experiments.

Small RNAs as biomarkers to differentiate benign and malign prostate diseases: An alternative for transrectal punch biopsy of the prostate?

Prostate cancer (PCa) is the most common cancer and the third most frequent cause of male cancer death in Germany. MicroRNAs (miRNA) appear to be involved in the development and progression of PCa. A diagnostic differentiation from benign prostate hyperplasia (BPH) is often only possible through transrectal punch biopsy. This procedure is described as painful and carries risks. It was investigated whether urinary miRNAs can be used as biomarkers to differentiate the prostate diseases above. Therefore urine samples from urological patients with BPH (25) or PCa (28) were analysed using Next-Generation Sequencing to detect the expression profile of total and exosomal miRNA/piRNA. 79 miRNAs and 5 piwi-interacting RNAs (piRNAs) were significantly differentially expressed (adjusted p-value < 0.05 and log2-Fc > 1 or < -1). Of these, 6 miRNAs and 2 piRNAs could be statistically validated (AUC on test cohort > = 0.7). In addition, machine-learning algorithms were used to identify a panel of 22 additional miRNAs, whose interaction makes it possible to differentiate the groups as well. There are promising individual candidates for potential use as biomarkers in prostate cancer. The innovative approach of applying machine learning methods to this kind of data could lead to further small RNAs coming into scientific focus, which have so far been neglected.

Implementing FAIR data management within the German Network for Bioinformatics Infrastructure (de.NBI) exemplified by selected use cases.

This article describes some use case studies and self-assessments of FAIR status of de.NBI services to illustrate the challenges and requirements for the definition of the needs of adhering to the FAIR (findable, accessible, interoperable and reusable) data principles in a large distributed bioinformatics infrastructure. We address the challenge of heterogeneity of wet lab technologies, data, metadata, software, computational workflows and the levels of implementation and monitoring of FAIR principles within the different bioinformatics sub-disciplines joint in de.NBI. On the one hand, this broad service landscape and the excellent network of experts are a strong basis for the development of useful research data management plans. On the other hand, the large number of tools and techniques maintained by distributed teams renders FAIR compliance challenging.

BIONDA: a free database for a fast information on published biomarkers.

Summary: Because of the steadily increasing and already manually unmanageable total number of biomarker-related articles in biomedical research, there is a need for intelligent systems that extract all relevant information from biomedical texts and provide it as structured information to researchers in a user-friendly way. To address this, BIONDA was implemented as a free text mining-based online database for molecular biomarkers including genes, proteins and miRNAs and for all kinds of diseases. The contained structured information on published biomarkers is extracted automatically from Europe PMC publication abstracts and high-quality sources like UniProt and Disease Ontology. This allows frequent content updates. Availability and implementation: BIONDA is freely accessible via a user-friendly web application at http://bionda.mpc.ruhr-uni-bochum.de. The current BIONDA code is available at GitHub via https://github.com/mpc-bioinformatics/bionda. Supplementary information: Supplementary data are available at Bioinformatics Advances online.

Systematic Comparison of Label-Free, SILAC, and TMT Techniques to Study Early Adaption toward Inhibition of EGFR Signaling in the Colorectal Cancer Cell Line DiFi.

We evaluated the quantification strategies label-free (LF), stable isotope labeling by amino acids in cell culture (SILAC), and tandem mass tags (TMT) and their performance in quantification of proteins and phosphosites (p-sites) to identify the most powerful approach for monitoring cellular signaling. We analyzed the epidermal growth factor receptor (EGFR) signaling network, which plays an essential role in colorectal cancer, and studied its dynamics within 24 h upon treatment with the EGFR-blocking antibody cetuximab, representing the first cellular adaption toward therapy. LF achieved superior coverage but was outperformed by label-based approaches regarding technical variability, especially for quantification of p-sites. TMT showed the lowest coverage and most missing values. We found that its performance considerably decreases when experimental replicates are distributed over several TMT plexes. SILAC showed the highest precision and outstanding performance for quantification of p-sites, rendering it the method of choice for analyzing cellular signaling in cell culture models. On the protein level, we observed only little regulation upon cetuximab treatment, whereas a great fraction of p-sites was significantly regulated. These dynamics represented an initial downregulation of the MAPK pathway, which was partially rescued as early as 24 h after treatment. We identified upregulation and signaling via ERBB3 as well as calcium and cAMP signaling as possible mechanisms bypassing the blockage of EGFR.

An intrinsically disordered proteins community for ELIXIR.

Intrinsically disordered proteins (IDPs) and intrinsically disordered regions (IDRs) are now recognised as major determinants in cellular regulation. This white paper presents a roadmap for future e-infrastructure developments in the field of IDP research within the ELIXIR framework. The goal of these developments is to drive the creation of high-quality tools and resources to support the identification, analysis and functional characterisation of IDPs. The roadmap is the result of a workshop titled "An intrinsically disordered protein user community proposal for ELIXIR" held at the University of Padua. The workshop, and further consultation with the members of the wider IDP community, identified the key priority areas for the roadmap including the development of standards for data annotation, storage and dissemination; integration of IDP data into the ELIXIR Core Data Resources; and the creation of benchmarking criteria for IDP-related software. Here, we discuss these areas of priority, how they can be implemented in cooperation with the ELIXIR platforms, and their connections to existing ELIXIR Communities and international consortia. The article provides a preliminary blueprint for an IDP Community in ELIXIR and is an appeal to identify and involve new stakeholders.

Digital droplet PCR-based chimerism analysis for monitoring of hematopoietic engraftment after allogeneic stem cell transplantation.

INTRODUCTION: Allogeneic hematopoietic stem cell transplantation (alloHSCT) is a curative approach for multiple hematologic diseases. The success of alloHSCT is evaluated by analyzing the proportion of living donor cells in blood and bone marrow samples of the recipient (chimerism analysis). To monitor the engrafted cells, donor's individual genetic markers are analyzed in peripheral blood and bone marrow samples, usually by using short tandem repeat (STR) analysis. An alternative method to measure chimerism is based on insertion and deletion markers (InDels) analyzed by digital droplet PCR (ddPCR); however, this approach is rarely evaluated in clinical practice. METHODS: In this study, we examined the usefulness of ddPCR-based chimerism analysis against the standard STR analysis in samples around day+30 after alloHSCT in clinical practice using peripheral blood and bone marrow samples. RESULTS: The median absolute difference between ddPCR and STR analysis was 0.55% points for bone marrow chimerisms and 0.25% points for peripheral blood chimerisms, respectively, including variation in the range of maximum 2% for both methods. The results of every single sample gave the same clinical message. CONCLUSION: According to our data, chimerism analysis by ddPCR has an excellent correlation with STR-based analyses. Due to its fast and easy applicability, the ddPCR technique is suitable for chimerism monitoring in clinical practice.

Monocytes-neutrophils-ratio as predictive marker for failure of first induction therapy in AML.

INTRODUCTION: Acute myeloid leukemia (AML) is, if untreated, a fatal hematologic neoplasia. Failure of the first induction chemotherapy is a hallmark for a poor prognosis. Early recognition of therapy failure is crucial for planning further therapies. Therefore, international guidelines recommend a bone marrow biopsy around day 14 after the beginning of induction therapy. Hypocellular bone marrow on day 14 is still gold standard for therapy assessment and further therapy strategy. Despite this, non-invasive ways for the evaluation of induction therapy were looked for in the past years. METHODS: We collected peripheral blood cell counts and routine laboratory values of patients treated with "7 + 3" induction therapy. Ratios of absolute cell counts of monocytes and neutrophils (MNR) were calculated daily, and the values were compared in patients with failure of the first induction therapy and patients with therapy response. RESULTS: 54 patients were included, 12 of which had failure of first induction therapy. The MNR following therapy was highly correlated with the bone marrow results. With the right cut-off, the MNR provides a valid and reliable tool for identification of patients with failure of first induction therapy with a sensitivity of 83.3% and a specificity of 87.8% on day 18. CONCLUSIONS: We propose a novel and non-invasive method for detection of failure of first induction therapy in patients with de novo AML and "7 + 3" induction therapy. The MNR is free of cost since the required cell counts are performed routinely for each patient undergoing intensive chemotherapy.

Protein variability in cerebrospinal fluid and its possible implications for neurological protein biomarker research.

Cerebrospinal fluid is investigated in biomarker studies for various neurological disorders of the central nervous system due to its proximity to the brain. Currently, only a limited number of biomarkers have been validated in independent studies. The high variability in the protein composition and protein abundance of cerebrospinal fluid between as well as within individuals might be an important reason for this phenomenon. To evaluate this possibility, we investigated the inter- and intraindividual variability in the cerebrospinal fluid proteome globally, with a specific focus on disease biomarkers described in the literature. Cerebrospinal fluid from a longitudinal study group including 12 healthy control subjects was analyzed by label-free quantification (LFQ) via LC-MS/MS. Data were quantified via MaxQuant. Then, the intra- and interindividual variability and the reference change value were calculated for every protein. We identified and quantified 791 proteins, and 216 of these proteins were abundant in all samples and were selected for further analysis. For these proteins, we found an interindividual coefficient of variation of up to 101.5% and an intraindividual coefficient of variation of up to 29.3%. Remarkably, these values were comparably high for both proteins that were published as disease biomarkers and other proteins. Our results support the hypothesis that natural variability greatly impacts cerebrospinal fluid protein biomarkers because high variability can lead to unreliable results. Thus, we suggest controlling the variability of each protein to distinguish between good and bad biomarker candidates, e.g., by utilizing reference change values to improve the process of evaluating potential biomarkers in future studies.

MicroRNA-30c as a novel diagnostic biomarker for primary and secondary B-cell lymphoma of the CNS.

Primary lymphomas of the central nervous system (PCNSL) are highly aggressive tumors affecting exclusively the CNS, meninges, and eyes. PCNSL must be separated from secondary spread of systemic lymphoma to the CNS (SCNSL), which may occur at diagnosis or relapse of systemic lymphomas. At present, there are no valid methods to distinguish PCNSL from SCNSL based on tumor biopsy because of similar histological presentation. However, SCNSL and PCNSL are different in terms of prognosis and adequate therapy protocols. MicroRNA expression profiles of CSF samples collected from SCNSL and PCNSL patients were compared using microRNA arrays. MiR-30c revealed the largest differential expression and was selected for validation by RT-PCR on 61 CSF samples from patients with PCNSL and 14 samples from SCNSL. MiR-30c was significantly increased in patients with SCNSL compared to PCNSL (p < 0.001). MiR-30c levels in CSF enabled the differentiation of patients with PCNSL from SCNSL with an area under the curve (AUC) of 0.86, with a sensitivity of 90.9% and a specificity of 85.5%. Our data suggest that miR-30c detected in the CSF can serve as biomarker for distinction between PCNSL and SCNSL. The validation in a larger cohort is needed. With respect to its function, miR-30c may facilitate lymphoma cells to engraft into CNS by interaction with CELSR3 gene that controls the function of ependymal cilia and, thus, affects the circulation of CSF.

BioInfra.Prot: A comprehensive proteomics workflow including data standardization, protein inference, expression analysis and data publication.

The analysis of high-throughput mass spectrometry-based proteomics data must address the specific challenges of this technology. To this end, the comprehensive proteomics workflow offered by the de.NBI service center BioInfra.Prot provides indispensable components for the computational and statistical analysis of this kind of data. These components include tools and methods for spectrum identification and protein inference, protein quantification, expression analysis as well as data standardization and data publication. All particular methods of the workflow which address these tasks are state-of-the-art or cutting edge. As has been shown in previous publications, each of these methods is adequate to solve its specific task and gives competitive results. However, the methods included in the workflow are continuously reviewed, updated and improved to adapt to new scientific developments. All of these particular components and methods are available as stand-alone BioInfra.Prot services or as a complete workflow. Since BioInfra.Prot provides manifold fast communication channels to get access to all components of the workflow (e.g., via the BioInfra.Prot ticket system: bioinfraprot@rub.de) users can easily benefit from this service and get support by experts.

Quantitative proteome analysis reveals the correlation between endocytosis-associated proteins and hepatocellular carcinoma dedifferentiation.

The majority of poorly differentiated hepatocellular carcinomas (HCCs) develop from well-differentiated tumors. Endocytosis is a cellular function which is likely to take part in this development due to its important role in regulating the abundances of vital signaling receptors. Here, we aimed to investigate the abundance of endocytosis-associated proteins in HCCs with various differentiation grades. Therefore, we analyzed 36 tissue specimens from HCC patients via LC-MS/MS-based label-free quantitative proteomics including 19 HCC tissue samples with different degrees of histological grades and corresponding non-tumorous tissue controls. As a result, 277 proteins were differentially regulated between well-differentiated tumors and controls. In moderately and poorly differentiated tumors, 278 and 1181 proteins, respectively, were significantly differentially regulated compared to non-tumorous tissue. We explored the regulated proteins based on their functions and identified thirty endocytosis-associated proteins, mostly overexpressed in poorly differentiated tumors. These included proteins that have been shown to be up-regulated in HCC like clathrin heavy chain-1 (CLTC) as well as unknown proteins, such as secretory carrier-associated membrane protein 3 (SCAMP3). The abundances of SCAMP3 and CLTC were immunohistochemically examined in tissue sections of 84 HCC patients. We demonstrate the novel association of several endocytosis-associated proteins, in particular, SCAMP3 with HCC progression.

PAA: an R/bioconductor package for biomarker discovery with protein microarrays.

UNLABELLED: The R/Bioconductor package Protein Array Analyzer (PAA) facilitates a flexible analysis of protein microarrays for biomarker discovery (esp., ProtoArrays). It provides a complete data analysis workflow including preprocessing and quality control, uni- and multivariate feature selection as well as several different plots and results tables to outline and evaluate the analysis results. As a main feature, PAA's multivariate feature selection methods are based on recursive feature elimination (e.g. SVM-recursive feature elimination, SVM-RFE) with stability ensuring strategies such as ensemble feature selection. This enables PAA to detect stable and reliable biomarker candidate panels. AVAILABILITY AND IMPLEMENTATION: PAA is freely available (BSD 3-clause license) from http://www.bioconductor.org/packages/PAA/ CONTACT: michael.turewicz@rub.de or martin.eisenacher@rub.de.