Gene expression and copy number profiling of follicular lymphoma biopsies from patients treated with first-line rituximab without chemotherapy

The Nordic Lymphoma Study Group has performed two randomized clinical trials with chemotherapy-free first-line treatment (rituximab +/- interferon) in follicular lymphoma (FL), with 73% of patients alive and 38% without any need of chemotherapy after 10.6 years median follow-up. In order to identify predictive markers, that may also serve as therapeutic targets, gene expression- and copy number profiles were obtained from 97 FL patients using whole genome microarrays. Copy number alterations (CNAs) were identified, e.g. by GISTIC. Cox Lasso Regression and Lasso logistic regression were used to determine molecular features predictive of time to next therapy (TTNT). A few molecular changes were associated with TTNT (e.g. increased expression of INPP5B, gains in 12q23/q24), but were not significant after adjusting for multiple testing. Our findings suggest that there are no strong determinants of patient outcome with respect to GE data and CNAs in FL patients treated with a chemotherapy-free regimen (i.e. rituximab +/- interferon).

Expectation of clinical decision support systems: a survey study among nephrologist end-users

BACKGROUND

Chronic kidney disease (CKD), a major public health problem with differing disease etiologies, leads to complications, comorbidities, polypharmacy, and mortality. Monitoring disease progression and personalized treatment efforts are crucial for long-term patient outcomes. Physicians need to integrate different data levels, e.g., clinical parameters, biomarkers, and drug information, with medical knowledge. Clinical decision support systems (CDSS) can tackle these issues and improve patient management. Knowledge about the awareness and implementation of CDSS in Germany within the field of nephrology is scarce.

PURPOSE

Nephrologists' attitude towards any CDSS and potential CDSS features of interest, like adverse event prediction algorithms, is important for a successful implementation. This survey investigates nephrologists' experiences with and expectations towards a useful CDSS for daily medical routine in the outpatient setting.

METHODS

The 38-item questionnaire survey was conducted either by telephone or as a do-it-yourself online interview amongst nephrologists across all of Germany. Answers were collected and analysed using the Electronic Data Capture System REDCap, as well as Stata SE 15.1, and Excel. The survey consisted of four modules: experiences with CDSS (M1), expectations towards a helpful CDSS (M2), evaluation of adverse event prediction algorithms (M3), and ethical aspects of CDSS (M4). Descriptive statistical analyses of all questions were conducted.

RESULTS

The study population comprised 54 physicians, with a response rate of about 80-100% per question. Most participants were aged between 51-60 years (45.1%), 64% were male, and most participants had been working in nephrology out-patient clinics for a median of 10.5 years. Overall, CDSS use was poor (81.2%), often due to lack of knowledge about existing CDSS. Most participants (79%) believed CDSS to be helpful in the management of CKD patients with a high willingness to try out a CDSS. Of all adverse event prediction algorithms, prediction of CKD progression (97.8%) and in-silico simulations of disease progression when changing, e. g., lifestyle or medication (97.7%) were rated most important. The spectrum of answers on ethical aspects of CDSS was diverse.

CONCLUSION

This survey provides insights into experience with and expectations of out-patient nephrologists on CDSS. Despite the current lack of knowledge on CDSS, the willingness to integrate CDSS into daily patient care, and the need for adverse event prediction algorithms was high.

Anomaly detection in mixed high-dimensional molecular data

MOTIVATION

Mixed molecular data combines continuous and categorical features of the same samples, such as OMICS profiles with genotypes, diagnoses, or patient sex. Like all high-dimensional molecular data, it is prone to incorrect values that can stem from various sources for example the technical limitations of the measurement devices, errors in the sample preparation, or contamination. Most anomaly detection algorithms identify complete samples as outliers or anomalies. However, in most cases, not all measurements of those samples are erroneous but only a few one-dimensional features within the samples are incorrect. These one-dimensional data errors are continuous measurements that are either located outside or inside the normal ranges of their features but in both cases show atypical values given all other continuous and categorical features in the sample. Additionally, categorical anomalies can occur for example when the genotype or diagnosis was submitted wrongly.

RESULTS

We introduce ADMIRE (Anomaly Detection using MIxed gRaphical modEls), a novel approach for the detection and correction of anomalies in mixed high-dimensional data. Hereby, we focus on the detection of single (one-dimensional) data errors in the categorical and continuous features of a sample. For that the joint distribution of continuous and categorical features is learned by mixed graphical models, anomalies are detected by the difference between measured and model-based estimations and are corrected using imputation. We evaluated ADMIRE in simulation and by screening for anomalies in one of our own metabolic datasets. In simulation experiments, ADMIRE outperformed the state-of-the-art methods of Local Outlier Factor, stray, and Isolation Forest.

AVAILABILITY AND IMPLEMENTATION

All data and code is available at https://github.com/spang-lab/adadmire. ADMIRE is implemented in a Python package called adadmire which can be found at https://pypi.org/project/adadmire.

An overview and a roadmap for artificial intelligence in hematology and oncology

BACKGROUND

Artificial intelligence (AI) is influencing our society on many levels and has broad implications for the future practice of hematology and oncology. However, for many medical professionals and researchers, it often remains unclear what AI can and cannot do, and what are promising areas for a sensible application of AI in hematology and oncology. Finally, the limits and perils of using AI in oncology are not obvious to many healthcare professionals.

METHODS

In this article, we provide an expert-based consensus statement by the joint Working Group on "Artificial Intelligence in Hematology and Oncology" by the German Society of Hematology and Oncology (DGHO), the German Association for Medical Informatics, Biometry and Epidemiology (GMDS), and the Special Interest Group Digital Health of the German Informatics Society (GI). We provide a conceptual framework for AI in hematology and oncology.

RESULTS

First, we propose a technological definition, which we deliberately set in a narrow frame to mainly include the technical developments of the last ten years. Second, we present a taxonomy of clinically relevant AI systems, structured according to the type of clinical data they are used to analyze. Third, we show an overview of potential applications, including clinical, research, and educational environments with a focus on hematology and oncology.

CONCLUSION

Thus, this article provides a point of reference for hematologists and oncologists, and at the same time sets forth a framework for the further development and clinical deployment of AI in hematology and oncology in the future.

Treatment Strategies and Prognostic Factors in Secondary Central Nervous System Lymphoma: A Multicenter Study of 124 Patients

Secondary central nervous system lymphoma (SCNSL) is a rare and difficult to treat type of Non-Hodgkin lymphoma characterized by systemic and central nervous system (CNS) disease manifestations. In this study, 124 patients with SCNSL intensively treated and with clinical long-term follow-up were included. Initial histopathology, as divided in low-grade, other aggressive, and diffuse large B-cell lymphoma (DLBCL), was of prognostic significance. Overall response to induction treatment was a prognostic factor with early responding DLBCL-SCNSL in comparison to those non-responding experiencing a significantly better progression-free survival (PFS) and overall survival (OS). However, the type of induction regime was not prognostic for survival. Following consolidating high-dose chemotherapy and autologous stem cell transplantation (HDT-ASCT), DLBCL-SCNSL patients had better median PFS and OS. The important role of HDT-ASCT was further highlighted by favorable responses and survival of patients not responding to induction therapy and by excellent results in patients with de novo DLBCL-SCNSL (65% long-term survival). SCNSL identified as a progression of disease within 6 months of initial systemic lymphoma presentation represented a previously not appreciated subgroup with particularly dismal outcome. This temporal stratification model of SCNSL diagnosis revealed CNS progression of disease within 6 months as a promising candidate prognosticator for future studies.

Lipid remodeling of contrasting maize (Zea mays L.) hybrids under repeated drought

The role of recovery after drought has been proposed to play a more prominent role during the whole drought-adaption process than previously thought. Two maize hybrids with comparable growth but contrasting physiological responses were investigated using physiological, metabolic, and lipidomic tools to understand the plants' strategies of lipid remodeling in response to repeated drought stimuli. Profound differences in adaptation between hybrids were discovered during the recovery phase, which likely gave rise to different degrees of lipid adaptability to the subsequent drought event. These differences in adaptability are visible in galactolipid metabolism and fatty acid saturation patterns during recovery and may lead to a membrane dysregulation in the sensitive maize hybrid. Moreover, the more drought-tolerant hybrid displays more changes of metabolite and lipid abundance with a higher number of differences within individual lipids, despite a lower physiological response, while the responses in the sensitive hybrid are higher in magnitude but lower in significance on the level of individual lipids and metabolites. This study suggests that lipid remodeling during recovery plays a key role in the drought response of plants.

The Network Zoo: a multilingual package for the inference and analysis of gene regulatory networks

Inference and analysis of gene regulatory networks (GRNs) require software that integrates multi-omic data from various sources. The Network Zoo (netZoo; netzoo.github.io) is a collection of open-source methods to infer GRNs, conduct differential network analyses, estimate community structure, and explore the transitions between biological states. The netZoo builds on our ongoing development of network methods, harmonizing the implementations in various computing languages and between methods to allow better integration of these tools into analytical pipelines. We demonstrate the utility using multi-omic data from the Cancer Cell Line Encyclopedia. We will continue to expand the netZoo to incorporate additional methods.

DRAGON: Determining Regulatory Associations using Graphical models on multi-Omic Networks

The increasing quantity of multi-omic data, such as methylomic and transcriptomic profiles collected on the same specimen or even on the same cell, provides a unique opportunity to explore the complex interactions that define cell phenotype and govern cellular responses to perturbations. We propose a network approach based on Gaussian Graphical Models (GGMs) that facilitates the joint analysis of paired omics data. This method, called DRAGON (Determining Regulatory Associations using Graphical models on multi-Omic Networks), calibrates its parameters to achieve an optimal trade-off between the network's complexity and estimation accuracy, while explicitly accounting for the characteristics of each of the assessed omics 'layers.' In simulation studies, we show that DRAGON adapts to edge density and feature size differences between omics layers, improving model inference and edge recovery compared to state-of-the-art methods. We further demonstrate in an analysis of joint transcriptome - methylome data from TCGA breast cancer specimens that DRAGON can identify key molecular mechanisms such as gene regulation via promoter methylation. In particular, we identify Transcription Factor AP-2 Beta (TFAP2B) as a potential multi-omic biomarker for basal-type breast cancer. DRAGON is available as open-source code in Python through the Network Zoo package (netZooPy v0.8; netzoo.github.io).

MO474: Expectation and Acceptance of a Clinical Decision Support Software by Nephrologist End-Users: The Ckdnapp Survey

BACKGROUND AND AIMS

Chronic kidney disease (CKD) is a major public health problem. CKD constitutes a complex disease due to differing underlying disease etiologies in each patient, which can in turn lead to many complications, comorbidities and polypharmacy. Monitoring disease progression and personalized treatment efforts are crucial for optimal long-term patient outcomes. In order to achieve this, physicians need to integrate different levels of data, e.g. clinical/demographic parameters, biomarkers and drug information, with medical knowledge. Clinical decision support systems (CDSS) can tackle these issues and improve patient management. ‘CKDNapp’ (CKD Nephrologist App), a CDSS application for nephrologists, based on mathematical models using machine-learning techniques, is currently being developed (https://ckdn.app). CKDNapp is intended to become a tool for daily clinical use. The nephrologists’ attitude towards any CDSS and CKDNapp in particular is of prime importance for its successful implementation into the daily medical routine. This survey investigates nephrologists’ experiences with CDSS in general and their expectations towards a reliable and useful application supporting their daily medical routine.

METHOD

CKDNapp survey is ongoing and has been conducted by telephone or as a do-it-yourself online interview in the form of a 38-item questionnaire. The answers of nephrologists from all regions across Germany were collected and analyzed using the Electronic Data Capture System, RedCap [1]. CKDNapp survey is divided into four modules: (1) experiences with CDSS, (2) expectations of a helpful CDSS, (3) evaluation of the planned contents of CKDNapp and (4) ethical aspects of CDSS (in collaboration with the BMBF-funded DESIREE project; https://www.desiree-forschung.de/desiree/index.php). All questions were based on a literature search for questionnaire items on CDSS [2, 3]. Response formats include the Likert scale or multiple choice. Descriptive statistical analyses of all questions were calculated.

RESULTS

In total, 44 participants took the survey, and completeness of answers ranged from 85% to 100%. Participants were aged 51–60 years old, male (64%) and had been working in nephrology outpatient clinics for a median of 12 years. Nephrologists treated a median of 35 patients/day. A total of 85% of participants reported never or rarely use a CDSS in patient care. The most frequently given reason for this was a lack of knowledge about CDSS. Nevertheless, 79% of participants believed CDSS to be helpful in the management of patients with CKD and 71% would be willing to use a CDSS given the chance to do so. When rating the importance of planned CKDNapp features, prediction of CKD progression (97%, Figure 1) and in-silico simulations of disease progression when changing, e.g. lifestyle or medication (97%) were most important, followed by the need for integration of available CKD guidelines (95%), prediction of acute kidney injury (95%), prediction of mortality risk (80%), and easy access to patient information (76%). The spectrum of answers to ethical aspects of CDSS (utility of CDSS for experienced versus inexperienced nephrologists, aspects of machine learning Fig. 2, discrimination of minority groups, etc.) was diverse.

CONCLUSION

This survey provides insights into experience with and expectations of outpatient nephrologists on CDSS in general and CKDNapp in particular. Despite the current lack of knowledge on CDSS, the willingness to integrate CDSS into daily patient care and the evaluation of planned CKDNapp features was high.

OUP accepted manuscript

MOTIVATION

Estimating the effects of interventions on patient outcome is one of the key aspects of personalized medicine. Their inference is often challenged by the fact that the training data comprises only the outcome for the administered treatment, and not for alternative treatments (the so-called counterfactual outcomes). Several methods were suggested for this scenario based on observational data, i.e. data where the intervention was not applied randomly, for both continuous and binary outcome variables. However, patient outcome is often recorded in terms of time-to-event data, comprising right-censored event times if an event does not occur within the observation period. Albeit their enormous importance, time-to-event data are rarely used for treatment optimization. We suggest an approach named BITES (Balanced Individual Treatment Effect for Survival data), which combines a treatment-specific semi-parametric Cox loss with a treatment-balanced deep neural network; i.e. we regularize differences between treated and non-treated patients using Integral Probability Metrics (IPM).

RESULTS

We show in simulation studies that this approach outperforms the state of the art. Furthermore, we demonstrate in an application to a cohort of breast cancer patients that hormone treatment can be optimized based on six routine parameters. We successfully validated this finding in an independent cohort.

AVAILABILITY AND IMPLEMENTATION

We provide BITES as an easy-to-use python implementation including scheduled hyper-parameter optimization (https://github.com/sschrod/BITES). The data underlying this article are available in the CRAN repository at https://rdrr.io/cran/survival/man/gbsg.html and https://rdrr.io/cran/survival/man/rotterdam.html.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Chronic Kidney Disease Cohort Studies: A Guide to Metabolome Analyses

Kidney diseases still pose one of the biggest challenges for global health, and their heterogeneity and often high comorbidity load seriously hinders the unraveling of their underlying pathomechanisms and the delivery of optimal patient care. Metabolomics, the quantitative study of small organic compounds, called metabolites, in a biological specimen, is gaining more and more importance in nephrology research. Conducting a metabolomics study in human kidney disease cohorts, however, requires thorough knowledge about the key workflow steps: study planning, sample collection, metabolomics data acquisition and preprocessing, statistical/bioinformatics data analysis, and results interpretation within a biomedical context. This review provides a guide for future metabolomics studies in human kidney disease cohorts. We will offer an overview of important a priori considerations for metabolomics cohort studies, available analytical as well as statistical/bioinformatics data analysis techniques, and subsequent interpretation of metabolic findings. We will further point out potential research questions for metabolomics studies in the context of kidney diseases and summarize the main results and data availability of important studies already conducted in this field.

A Predictive Model for Progression of CKD to Kidney Failure Based on Routine Laboratory Tests

RATIONALE & OBJECTIVE

Stratification of chronic kidney disease (CKD) patients at risk for progressing to end-stage kidney disease (ESKD) requiring kidney replacement therapy (KRT) is important for clinical decision-making and trial enrollment.

STUDY DESIGN

Four independent prospective observational cohort studies.

SETTING & PARTICIPANTS

The development cohort was comprised of 4,915 CKD patients and three independent validation cohorts were comprised of a total of 3,063. Patients were followed-up for approximately five years.

NEW PREDICTORS & ESTABLISHED PREDICTORS

22 demographic, anthropometric and laboratory variables commonly assessed in CKD patients.

OUTCOMES

Progression to ESKD requiring KRT.

ANALYTICAL APPROACH

A Least Absolute Shrinkage and Selection Operator (LASSO) Cox proportional hazards model was fit to select laboratory variables that best identified patients at high risk for ESKD. Model discrimination and calibration were assessed and compared against the 4-variable Tangri (T4) risk equation. Both used a resampling approach within the development cohort and in the validation cohorts using cause-specific concordance (C) statistics, net reclassification improvement, and calibration graphs.

RESULTS

The newly derived 6-variable (Z6) risk score included serum creatinine, albumin, cystatin C and urea, as well as hemoglobin and the urine albumin-to-creatinine ratio. Based on the resampling approach, Z6 achieved a median C value of 0.909 (95% CI, 0.868-0.937) at two years after the baseline visit, whereas the T4 achieved a median C value of 0.855 (95% CI, 0.799-0.915). In the three independent validation cohorts, Z6 C values were 0.894, 0.921, and 0.891, whereas the T4 C values were 0.882, 0.913, and 0.862.

LIMITATIONS

The Z6 was both derived and tested only in White European cohorts.

CONCLUSIONS

A new risk equation, based on six routinely available laboratory tests facilitates identification of patients with CKD who are at high risk of progressing to ESKD.

Proteome profiling of repeated drought stress reveals genotype-specific responses and memory effects in maize

Drought has become a major stress for agricultural productivity in temperate regions, such as central Europe. Thus, information on how crop plants respond to drought is important to develop tolerant hybrids and to ensure yield stability. Posttranscriptional regulation through changed protein abundances is an important mechanism of short-term response to stress events that has not yet been widely exploited in breeding strategies. Here, we investigated the response to repeated drought exposure of a tolerant and a sensitive maize hybrid in order to understand general protein abundance changes induced by singular drought or repeated drought events. In general, drought affected protein abundance of multiple pathways in the plant. We identified starch metabolism, aquaporin abundance, PSII proteins and histones as strongly associated with typical drought-induced phenotypes such as increased membrane leakage, osmolality or effects on stomatal conductance and assimilation rate. In addition, we found a strong effect of drought on nutrient assimilation, especially the sulfur metabolism. In general, pre-experience of mild drought before exposure to a more severe drought resulted in visible adaptations resulting in dampened phenotypes as well as lower magnitude of protein abundance changes.

Identification of a miRNA based model to detect prognostic subgroups in patients with aggressive B-cell lymphoma

In order to differentiate prognostic subgroups of patients with aggressive B-cell lymphoma, we analyzed the expression of 800 miRNAs with the NanoString nCounter human miRNA assay on a cohort of 228 FFPE samples of patients enrolled in the RICOVER-60 and MegaCHOEP trials. We identified significant miRNA signatures for overall survival (OS) and progression-free survival (PFS) by LASSO-penalized linear Cox-regression. High expression levels of miR-130a-3p and miR-423-5p indicate a better prognosis, whereas high levels of miR-374b-5p, miR-590-5p, miR-186-5p, and miR-106b-5p increase patients' risk levels for OS. Regarding PFS high expression of miR-365a-5p in addition to the other two miRNAs improves the prognosis and high levels of miR374a-5p, miR-106b-5p, and miR-590-5p, connects with increased risk and poor prognosis. We identified miRNA signatures to subdivide patients into two different risk groups. These prognostic models may be used in risk stratification in future clinical trials and help making personalized therapy decisions.

Gaussian and Mixed Graphical Models as (multi-)omics data analysis tools

Gaussian Graphical Models (GGMs) are tools to infer dependencies between biological variables. Popular applications are the reconstruction of gene, protein, and metabolite association networks. GGMs are an exploratory research tool that can be useful to discover interesting relations between genes (functional clusters) or to identify therapeutically interesting genes, but do not necessarily infer a network in the mechanistic sense. Although GGMs are well investigated from a theoretical and applied perspective, important extensions are not well known within the biological community. GGMs assume, for instance, multivariate normal distributed data. If this assumption is violated Mixed Graphical Models (MGMs) can be the better choice. In this review, we provide the theoretical foundations of GGMs, present extensions such as MGMs or multi-class GGMs, and illustrate how those methods can provide insight in biological mechanisms. We summarize several applications and present user-friendly estimation software. This article is part of a Special Issue entitled: Transcriptional Profiles and Regulatory Gene Networks edited by Dr. Dr. Federico Manuel Giorgi and Dr. Shaun Mahony.

DTD: An R Package for Digital Tissue Deconvolution

Digital tissue deconvolution (DTD) estimates the cellular composition of a tissue from its bulk gene-expression profile. For this, DTD approximates the bulk as a mixture of cell-specific expression profiles. Different tissues have different cellular compositions, with cells in different activation states, and embedded in different environments. Consequently, DTD can profit from tailoring the deconvolution model to a specific tissue context. Loss-function learning adapts DTD to a specific tissue context, such as the deconvolution of blood, or a specific type of tumor tissue. We provide software for loss-function learning, for its validation and visualization, and for applying the DTD models to new data.

Automated macrophage counting in DLBCL tissue samples: a ROF filter based approach

BACKGROUND

For analysis of the tumor microenvironment in diffuse large B-cell lymphoma (DLBCL) tissue samples, it is desirable to obtain information about counts and distribution of different macrophage subtypes. Until now, macrophage counts are mostly inferred from gene expression analysis of whole tissue sections, providing only indirect information. Direct analysis of immunohistochemically (IHC) fluorescence stained tissue samples is confronted with several difficulties, e.g. high variability of shape and size of target macrophages and strongly inhomogeneous intensity of staining. Consequently, application of commercial software is largely restricted to very rough analysis modes, and most macrophage counts are still obtained by manual counting in microarrays or high power fields, thus failing to represent the heterogeneity of tumor microenvironment adequately.

METHODS

We describe a Rudin-Osher-Fatemi (ROF) filter based segmentation approach for whole tissue samples, combining floating intensity thresholding and rule-based feature detection. Method is validated against manual counts and compared with two commercial software kits (Tissue Studio 64, Definiens AG, and Halo, Indica Labs) and a straightforward machine-learning approach in a set of 50 test images. Further, the novel method and both commercial packages are applied to a set of 44 whole tissue sections. Outputs are compared with gene expression data available for the same tissue samples. Finally, the ROF based method is applied to 44 expert-specified tumor subregions for testing selection and subsampling strategies.

RESULTS

Among all tested methods, the novel approach is best correlated with manual count (0.9297). Automated detection of evaluation subregions proved to be fully reliable. Comparison with gene expression data obtained for the same tissue samples reveals only moderate to low correlation levels. Subsampling within tumor subregions is possible with results almost identical to full sampling. Mean macrophage size in tumor subregions is 152.5±111.3 μm2.

CONCLUSIONS

ROF based approach is successfully applied to detection of IHC stained macrophages in DLBCL tissue samples. The method competes well with existing commercial software kits. In difference to them, it is fully automated, externally repeatable, independent on training data and completely documented. Comparison with gene expression data indicates that image morphometry constitutes an independent source of information about antibody-polarized macrophage occurence and distribution.

A Novel Metabolic Signature To Predict the Requirement of Dialysis or Renal Transplantation in Patients with Chronic Kidney Disease

Identification of chronic kidney disease patients at risk of progressing to end-stage renal disease (ESRD) is essential for treatment decision-making and clinical trial design. Here, we explored whether proton nuclear magnetic resonance (NMR) spectroscopy of blood plasma improves the currently best performing kidney failure risk equation, the so-called Tangri score. Our study cohort comprised 4640 participants from the German Chronic Kidney Disease (GCKD) study, of whom 185 (3.99%) progressed over a mean observation time of 3.70 ± 0.88 years to ESRD requiring either dialysis or transplantation. The original four-variable Tangri risk equation yielded a C statistic of 0.863 (95% CI, 0.831-0.900). Upon inclusion of NMR features by state-of-the-art machine learning methods, the C statistic improved to 0.875 (95% CI, 0.850-0.911), thereby outperforming the Tangri score in 94 out of 100 subsampling rounds. Of the 24 NMR features included in the model, creatinine, high-density lipoprotein, valine, acetyl groups of glycoproteins, and Ca²⁺-EDTA carried the highest weights. In conclusion, proton NMR-based plasma fingerprinting improved markedly the detection of patients at risk of developing ESRD, thus enabling enhanced patient treatment.

Statistical Analysis of NMR Metabolic Fingerprints: Established Methods and Recent Advances

In this review, we summarize established and recent bioinformatic and statistical methods for the analysis of NMR-based metabolomics. Data analysis of NMR metabolic fingerprints exhibits several challenges, including unwanted biases, high dimensionality, and typically low sample numbers. Common analysis tasks comprise the identification of differential metabolites and the classification of specimens. However, analysis results strongly depend on the preprocessing of the data, and there is no consensus yet on how to remove unwanted biases and experimental variance prior to statistical analysis. Here, we first review established and new preprocessing protocols and illustrate their pros and cons, including different data normalizations and transformations. Second, we give a brief overview of state-of-the-art statistical analysis in NMR-based metabolomics. Finally, we discuss a recent development in statistical data analysis, where data normalization becomes obsolete. This method, called zero-sum regression, builds metabolite signatures whose estimation as well as predictions are independent of prior normalization.

Principles of Systems Biology, No. 31

This month: selected work from the 2018 RECOMB meeting, organized by Ecole Polytechnique and held last April in Paris.

Round-robin test for the cell-of-origin classification of diffuse large B-cell lymphoma-a feasibility study using full slide staining

Diffuse large B-cell lymphoma (DLBCL) is subdivided by gene expression analysis (GEP) into two molecular subtypes named germinal center B-cell-like (GCB) and activated B-cell-like (ABC) after their putative cell-of-origin (COO). Determination of the COO is considered mandatory in any new-diagnosed DLBCL, not otherwise specified according to the updated WHO classification. Despite the fact that pathologists are free to choose the method for COO classification, immunohistochemical (IHC) assays are most widely used. However, to the best of our knowledge, no round-robin test to evaluate the interlaboratory variability has been published so far. Eight hematopathology laboratories participated in an interlaboratory test for COO classification of 10 DLBCL tumors using the IHC classifier comprising the expression of CD10, BCL6, and MUM1 (so-called Hans classifier). The results were compared with GEP for COO signature and, in a subset, with results obtained by image analysis. In 7/10 cases (70%), at least seven laboratories assigned a given case to the same COO subtype (one center assessed one sample as not analyzable), which was in agreement with the COO subtype determined by GEP. The results in 3/10 cases (30%) revealed discrepancies between centers and/or between IHC and GEP subtype. Whereas the CD10 staining results were highly reproducible, staining for MUM1 was inconsistent in 50% and for BCL6 in 40% of cases. Image analysis of 16 slides stained for BCL6 (N = 8) and MUM1 (N = 8) of the two cases with the highest disagreement in COO classification were in line with the score of the pathologists in 14/16 stainings analyzed (87.5%). This study describes the first round-robin test for COO subtyping in DLBCL using IHC and demonstrates that COO classification using the Hans classifier yields consistent results among experienced hematopathologists, even when variable staining protocols are used. Data from this small feasibility study need to be validated in larger cohorts.

Molecular signatures that can be transferred across different omics platforms

Motivation

Molecular signatures for treatment recommendations are well researched. Still it is challenging to apply them to data generated by different protocols or technical platforms.

Results

We analyzed paired data for the same tumors (Burkitt lymphoma, diffuse large B-cell lymphoma) and features that had been generated by different experimental protocols and analytical platforms including the nanoString nCounter and Affymetrix Gene Chip transcriptomics as well as the SWATH and SRM proteomics platforms. A statistical model that assumes independent sample and feature effects accounted for 69–94% of technical variability. We analyzed how variability is propagated through linear signatures possibly affecting predictions and treatment recommendations. Linear signatures with feature weights adding to zero were substantially more robust than unbalanced signatures. They yielded consistent predictions across data from different platforms, both for transcriptomics and proteomics data. Similarly stable were their predictions across data from fresh frozen and matching formalin-fixed paraffin-embedded human tumor tissue.

Availability and Implementation

The R-package ‘zeroSum’ can be downloaded at https://github.com/rehbergT/zeroSum. Complete data and R codes necessary to reproduce all our results can be received from the authors upon request.

Scale-Invariant Biomarker Discovery in Urine and Plasma Metabolite Fingerprints

Metabolomics data is typically scaled to a common reference like a constant volume of body fluid, a constant creatinine level, or a constant area under the spectrum. Such scaling of the data, however, may affect the selection of biomarkers and the biological interpretation of results in unforeseen ways. Here, we studied how both the outcome of hypothesis tests for differential metabolite concentration and the screening for multivariate metabolite signatures are affected by the choice of scale. To overcome this problem for metabolite signatures and to establish a scale-invariant biomarker discovery algorithm, we extended linear zero-sum regression to the logistic regression framework and showed in two applications to ¹H NMR-based metabolomics data how this approach overcomes the scaling problem. Logistic zero-sum regression is available as an R package as well as a high-performance computing implementation that can be downloaded at https://github.com/rehbergT/zeroSum .

Reference point insensitive molecular data analysis

MOTIVATION

In biomedicine, every molecular measurement is relative to a reference point, like a fixed aliquot of RNA extracted from a tissue, a defined number of blood cells, or a defined volume of biofluid. Reference points are often chosen for practical reasons. For example, we might want to assess the metabolome of a diseased organ but can only measure metabolites in blood or urine. In this case, the observable data only indirectly reflects the disease state. The statistical implications of these discrepancies in reference points have not yet been discussed.

RESULTS

Here, we show that reference point discrepancies compromise the performance of regression models like the LASSO. As an alternative, we suggest zero-sum regression for a reference point insensitive analysis. We show that zero-sum regression is superior to the LASSO in case of a poor choice of reference point both in simulations and in an application that integrates intestinal microbiome analysis with metabolomics. Moreover, we describe a novel coordinate descent based algorithm to fit zero-sum elastic nets.

AVAILABILITY AND IMPLEMENTATION

The R-package "zeroSum" can be downloaded at https://github.com/rehbergT/zeroSum Moreover, we provide all R-scripts and data used to produce the results of this manuscript as Supplementary Material CONTACT: Michael.Altenbuchinger@ukr.de, Thorsten.Rehberg@ukr.de and Rainer.Spang@ukr.deSupplementary information: Supplementary material is available at Bioinformatics online.