CODEX: COunterfactual Deep learning for the in silico EXploration of cancer cell line perturbations.

MOTIVATION: High-throughput screens (HTS) provide a powerful tool to decipher the causal effects of chemical and genetic perturbations on cancer cell lines. Their ability to evaluate a wide spectrum of interventions, from single drugs to intricate drug combinations and CRISPR-interference, has established them as an invaluable resource for the development of novel therapeutic approaches. Nevertheless, the combinatorial complexity of potential interventions makes a comprehensive exploration intractable. Hence, prioritizing interventions for further experimental investigation becomes of utmost importance. RESULTS: We propose CODEX (COunterfactual Deep learning for the in silico EXploration of cancer cell line perturbations) as a general framework for the causal modeling of HTS data, linking perturbations to their downstream consequences. CODEX relies on a stringent causal modeling strategy based on counterfactual reasoning. As such, CODEX predicts drug-specific cellular responses, comprising cell survival and molecular alterations, and facilitates the in silico exploration of drug combinations. This is achieved for both bulk and single-cell HTS. We further show that CODEX provides a rationale to explore complex genetic modifications from CRISPR-interference in silico in single cells. AVAILABILITY AND IMPLEMENTATION: Our implementation of CODEX is publicly available at https://github.com/sschrod/CODEX. All data used in this article are publicly available.

Adaptive digital tissue deconvolution.

MOTIVATION: The inference of cellular compositions from bulk and spatial transcriptomics data increasingly complements data analyses. Multiple computational approaches were suggested and recently, machine learning techniques were developed to systematically improve estimates. Such approaches allow to infer additional, less abundant cell types. However, they rely on training data which do not capture the full biological diversity encountered in transcriptomics analyses; data can contain cellular contributions not seen in the training data and as such, analyses can be biased or blurred. Thus, computational approaches have to deal with unknown, hidden contributions. Moreover, most methods are based on cellular archetypes which serve as a reference; e.g. a generic T-cell profile is used to infer the proportion of T-cells. It is well known that cells adapt their molecular phenotype to the environment and that pre-specified cell archetypes can distort the inference of cellular compositions. RESULTS: We propose Adaptive Digital Tissue Deconvolution (ADTD) to estimate cellular proportions of pre-selected cell types together with possibly unknown and hidden background contributions. Moreover, ADTD adapts prototypic reference profiles to the molecular environment of the cells, which further resolves cell-type specific gene regulation from bulk transcriptomics data. We verify this in simulation studies and demonstrate that ADTD improves existing approaches in estimating cellular compositions. In an application to bulk transcriptomics data from breast cancer patients, we demonstrate that ADTD provides insights into cell-type specific molecular differences between breast cancer subtypes. AVAILABILITY AND IMPLEMENTATION: A python implementation of ADTD and a tutorial are available at Gitlab and zenodo (doi:10.5281/zenodo.7548362).

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).

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.

Interactive exploration of adverse events and multimorbidity in CKD.

BACKGROUND AND HYPOTHESIS: Persons with chronic kidney disease (CKD) are at increased risk of adverse events, early mortality, and multimorbidity. A detailed overview of adverse event types and rates from a large CKD cohort under regular nephrological care is missing. We generated an interactive tool to enable exploration of adverse events and their combinations in the prospective, observational German CKD (GCKD) study. METHODS: The GCKD study enrolled 5217 participants under regular nephrological care with an estimated glomerular filtration rate of 30-60 or >60 mL/min/1.73m2 and an overt proteinuria. Cardio-, cerebro- and peripheral vascular, kidney, infection, and cancer events, as well as deaths were adjudicated following a standard operation procedure. We summarized these time-to-event data points for exploration in interactive graphs within an R shiny app. Multivariable adjusted Cox models for time to first event were fitted. Cumulative incidence functions, Kaplan-Meier curves and intersection plots were used to display main adverse events and their combinations by sex and CKD etiology. RESULTS: Over a median of 6.5 years, 10 271 events occurred in total and 680 participants (13.0%) died while 2947 participants (56.5%) experienced any event. The new publicly available interactive platform enables readers to scrutinize adverse events and their combinations as well as mortality trends as a gateway to better understand multimorbidity in CKD: incident rates per 1000 patient-years varied by event type, CKD etiology, and baseline characteristics. Incidence rates for the most frequent events and their recurrence were 113.6 (cardiovascular), 75.0 (kidney), and 66.0 (infection). Participants with diabetic kidney disease and men were more prone to experiencing events. CONCLUSION: This comprehensive explorative tool to visualize adverse events (https://gckd.diz.uk-erlangen.de/), their combination, mortality, and multimorbidity among persons with CKD may manifest as a valuable resource for patient care, identification of high-risk groups, health services, and public health policy planning.

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.

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 kidney failure requiring kidney replacement therapy (KFRT) is important for clinical decision-making and trial enrollment. STUDY DESIGN: Four independent prospective observational cohort studies. SETTING & PARTICIPANTS: The development cohort comprised 4,915 CKD patients, and 3 independent validation cohorts comprised a total of 3,063. Patients were observed for approximately 5 years. EXPOSURE: 22 demographic, anthropometric, and laboratory variables commonly assessed in CKD patients. OUTCOME: Progression to KFRT. 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 KFRT. Model discrimination and calibration were assessed and compared against the 4-variable Tangri (T4) risk equation both in 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 risk score (Z6) included serum creatinine, albumin, cystatin C, and urea, as well as hemoglobin and the urinary albumin-creatinine ratio. In the the resampling approach, Z6 achieved a median C statistic of 0.909 (95% CI, 0.868-0.937) at 2 years after the baseline visit, whereas the T4 achieved a median C statistic of 0.855 (95% CI, 0.799-0.915). In the 3 independent validation cohorts, the Z6C statistics were 0.894, 0.921, and 0.891, whereas the T4C statistics 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 6 routinely available laboratory tests facilitates identification of patients with CKD who are at high risk of progressing to KFRT.

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.

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.

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 .

Cell-of-origin classification by gene expression and MYC-rearrangements in diffuse large B-cell lymphoma of children and adolescents.

We present the largest series of diffuse large B-cell lymphoma (DLBCL) in patients younger than 18 years analysed to date by gene expression profiling using Nanostring technology to identify molecular subtypes and fluorescent in situ hybridization for translocations of MYC. We show that the activated B cell-like subtype of DLBCL is exceedingly rare in children and - in contrast to adults- not associated with outcome. Furthermore, we review the current literature and demonstrate that MYC translocations are not more frequent in paediatric compared to adult DLBCL. A prognostic role of MYC in the paediatric age groups seems unlikely.

Human leukocyte antigen (HLA) class I expression on Hodgkin-Reed-Sternberg cells is an EBV-independent major determinant of microenvironment composition in classic Hodgkin lymphoma.

Hodgkin-Reed-Sternberg cells (HRSCs) in classic Hodgkin Lymphoma (HL) frequently lack expression of human leukocyte antigen class I (HLA-I), considered to hamper activation of cytotoxic T cells in the tumor microenvironment (TME). Here, we demonstrate HLA-I expression on HRSCs to be a strong determinant of TME composition whereas expression of HLA-II was associated with only minor differential gene expression in the TME. In HLA-I-positive HL the HRSC content and expression of CCL17/TARC in HRSCs are low, independent of the presence of Epstein-Barr virus in HRSCs. Additionally, HLA-I-positive HL shows a high content of CD8+ cytotoxic T cells. However, an increased expression of the inhibitory immune checkpoint LAG3 on CD8+ T cells in close proximity to HRSCs is observed. Suggesting interference with cytotoxic activity, we observed an absence of clonally expanded T cells in the TME. While HLA-I-positive HL is not associated with an unfavorable clinical course in our cohorts, they share features with the recently described H2 subtype of HL. Given the major differences in TME composition, immune checkpoint inhibitors may differ in their mechanism of action in HLA-I-positive compared to HLA-I-negative HL.

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.

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.

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.

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.

Bucket Fuser: Statistical Signal Extraction for 1D 1H NMR Metabolomic Data.

Untargeted metabolomics is a promising tool for identifying novel disease biomarkers and unraveling underlying pathomechanisms. Nuclear magnetic resonance (NMR) spectroscopy is particularly suited for large-scale untargeted metabolomics studies due to its high reproducibility and cost effectiveness. Here, one-dimensional (1D) 1H NMR experiments offer good sensitivity at reasonable measurement times. Their subsequent data analysis requires sophisticated data preprocessing steps, including the extraction of NMR features corresponding to specific metabolites. We developed a novel 1D NMR feature extraction procedure, called Bucket Fuser (BF), which is based on a regularized regression framework with fused group LASSO terms. The performance of the BF procedure was demonstrated using three independent NMR datasets and was benchmarked against existing state-of-the-art NMR feature extraction methods. BF dynamically constructs NMR metabolite features, the widths of which can be adjusted via a regularization parameter. BF consistently improved metabolite signal extraction, as demonstrated by our correlation analyses with absolutely quantified metabolites. It also yielded a higher proportion of statistically significant metabolite features in our differential metabolite analyses. The BF algorithm is computationally efficient and it can deal with small sample sizes. In summary, the Bucket Fuser algorithm, which is available as a supplementary python code, facilitates the fast and dynamic extraction of 1D NMR signals for the improved detection of metabolic biomarkers.

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.