Scalable integration of multiomic single-cell data using generative adversarial networks.

MOTIVATION: Single-cell profiling has become a common practice to investigate the complexity of tissues, organs, and organisms. Recent technological advances are expanding our capabilities to profile various molecular layers beyond the transcriptome such as, but not limited to, the genome, the epigenome, and the proteome. Depending on the experimental procedure, these data can be obtained from separate assays or the very same cells. Yet, integration of more than two assays is currently not supported by the majority of the computational frameworks avaiable. RESULTS: We here propose a Multi-Omic data integration framework based on Wasserstein Generative Adversarial Networks suitable for the analysis of paired or unpaired data with a high number of modalities (>2). At the core of our strategy is a single network trained on all modalities together, limiting the computational burden when many molecular layers are evaluated. AVAILABILITY AND IMPLEMENTATION: Source code of our framework is available at https://github.com/vgiansanti/MOWGAN.

A Bayesian method to infer copy number clones from single-cell RNA and ATAC sequencing.

Single-cell RNA and ATAC sequencing technologies enable the examination of gene expression and chromatin accessibility in individual cells, providing insights into cellular phenotypes. In cancer research, it is important to consistently analyze these states within an evolutionary context on genetic clones. Here we present CONGAS+, a Bayesian model to map single-cell RNA and ATAC profiles onto the latent space of copy number clones. CONGAS+ clusters cells into tumour subclones with similar ploidy, rendering straightforward to compare their expression and chromatin profiles. The framework, implemented on GPU and tested on real and simulated data, scales to analyse seamlessly thousands of cells, demonstrating better performance than single-molecule models, and supporting new multi-omics assays. In prostate cancer, lymphoma and basal cell carcinoma, CONGAS+ successfully identifies complex subclonal architectures while providing a coherent mapping between ATAC and RNA, facilitating the study of genotype-phenotype maps and their connection to genomic instability.

Evolutionary signatures of human cancers revealed via genomic analysis of over 35,000 patients.

Recurring sequences of genomic alterations occurring across patients can highlight repeated evolutionary processes with significant implications for predicting cancer progression. Leveraging the ever-increasing availability of cancer omics data, here we unveil cancer's evolutionary signatures tied to distinct disease outcomes, representing "favored trajectories" of acquisition of driver mutations detected in patients with similar prognosis. We present a framework named ASCETIC (Agony-baSed Cancer EvoluTion InferenCe) to extract such signatures from sequencing experiments generated by different technologies such as bulk and single-cell sequencing data. We apply ASCETIC to (i) single-cell data from 146 myeloid malignancy patients and bulk sequencing from 366 acute myeloid leukemia patients, (ii) multi-region sequencing from 100 early-stage lung cancer patients, (iii) exome/genome data from 10,000+ Pan-Cancer Atlas samples, and (iv) targeted sequencing from 25,000+ MSK-MET metastatic patients, revealing subtype-specific single-nucleotide variant signatures associated with distinct prognostic clusters. Validations on several datasets underscore the robustness and generalizability of the extracted signatures.

LACE 2.0: an interactive R tool for the inference and visualization of longitudinal cancer evolution.

BACKGROUND: Longitudinal single-cell sequencing experiments of patient-derived models are increasingly employed to investigate cancer evolution. In this context, robust computational methods are needed to properly exploit the mutational profiles of single cells generated via variant calling, in order to reconstruct the evolutionary history of a tumor and characterize the impact of therapeutic strategies, such as the administration of drugs. To this end, we have recently developed the LACE framework for the Longitudinal Analysis of Cancer Evolution. RESULTS: The LACE 2.0 release aimed at inferring longitudinal clonal trees enhances the original framework with new key functionalities: an improved data management for preprocessing of standard variant calling data, a reworked inference engine, and direct connection to public databases. CONCLUSIONS: All of this is accessible through a new and interactive Shiny R graphical interface offering the possibility to apply filters helpful in discriminating relevant or potential driver mutations, set up inferential parameters, and visualize the results. The software is available at: github.com/BIMIB-DISCo/LACE.

SparseSignatures: An R package using LASSO-regularized non-negative matrix factorization to identify mutational signatures from human tumor samples.

We outline the features of the R package SparseSignatures and its application to determine the signatures contributing to mutation profiles of tumor samples. We describe installation details and illustrate a step-by-step approach to (1) prepare the data for signature analysis, (2) determine the optimal parameters, and (3) employ them to determine the signatures and related exposure levels in the point mutation dataset. For complete details on the use and execution of this protocol, please refer to Lal et al. (2021).

J-SPACE: a Julia package for the simulation of spatial models of cancer evolution and of sequencing experiments.

BACKGROUND: The combined effects of biological variability and measurement-related errors on cancer sequencing data remain largely unexplored. However, the spatio-temporal simulation of multi-cellular systems provides a powerful instrument to address this issue. In particular, efficient algorithmic frameworks are needed to overcome the harsh trade-off between scalability and expressivity, so to allow one to simulate both realistic cancer evolution scenarios and the related sequencing experiments, which can then be used to benchmark downstream bioinformatics methods. RESULT: We introduce a Julia package for SPAtial Cancer Evolution (J-SPACE), which allows one to model and simulate a broad set of experimental scenarios, phenomenological rules and sequencing settings.Specifically, J-SPACE simulates the spatial dynamics of cells as a continuous-time multi-type birth-death stochastic process on a arbitrary graph, employing different rules of interaction and an optimised Gillespie algorithm. The evolutionary dynamics of genomic alterations (single-nucleotide variants and indels) is simulated either under the Infinite Sites Assumption or several different substitution models, including one based on mutational signatures. After mimicking the spatial sampling of tumour cells, J-SPACE returns the related phylogenetic model, and allows one to generate synthetic reads from several Next-Generation Sequencing (NGS) platforms, via the ART read simulator. The results are finally returned in standard FASTA, FASTQ, SAM, ALN and Newick file formats. CONCLUSION: J-SPACE is designed to efficiently simulate the heterogeneous behaviour of a large number of cancer cells and produces a rich set of outputs. Our framework is useful to investigate the emergent spatial dynamics of cancer subpopulations, as well as to assess the impact of incomplete sampling and of experiment-specific errors. Importantly, the output of J-SPACE is designed to allow the performance assessment of downstream bioinformatics pipelines processing NGS data. J-SPACE is freely available at: https://github.com/BIMIB-DISCo/J-Space.jl .

Early detection and improved genomic surveillance of SARS-CoV-2 variants from deep sequencing data.

A key task of genomic surveillance of infectious viral diseases lies in the early detection of dangerous variants. Unexpected help to this end is provided by the analysis of deep sequencing data of viral samples, which are typically discarded after creating consensus sequences. Such analysis allows one to detect intra-host low-frequency mutations, which are a footprint of mutational processes underlying the origination of new variants. Their timely identification may improve public-health decision-making with respect to traditional approaches exploiting consensus sequences. We present the analysis of 220,788 high-quality deep sequencing SARS-CoV-2 samples, showing that many spike and nucleocapsid mutations of interest associated to the most circulating variants, including Beta, Delta, and Omicron, might have been intercepted several months in advance. Furthermore, we show that a refined genomic surveillance system leveraging deep sequencing data might allow one to pinpoint emerging mutation patterns, providing an automated data-driven support to virologists and epidemiologists.

Chemotherapy after PD-1 inhibitors in relapsed/refractory Hodgkin lymphoma: Outcomes and clonal evolution dynamics.

Checkpoint inhibitors (CPIs) are routinely employed in relapsed/refractory classical Hodgkin lymphoma. Nonetheless, persistent long-term responses are uncommon, and one-third of patients are refractory. Several reports have suggested that treatment with CPIs may re-sensitize patients to chemotherapy, however there is no consensus on the optimal chemotherapy regimen and subsequent consolidation strategy. In this retrospective study we analysed the response to rechallenge with chemotherapy after CPI failure. Furthermore, we exploratively characterized the clonal evolution profile of a small sample of patients (n = 5) by employing the CALDER approach. Among the 28 patients included in the study, 17 (71%) were primary refractory and 26 (92%) were refractory to the last chemotherapy prior to CPIs. Following rechallenge with chemotherapy, response was recorded in 23 (82%) patients experiencing complete remission and 3 (11%) patients experiencing partial remission. The tumour evolution of the patients inferred by CALDER seemingly occurred prior to the first cycle of therapy and was characterized either by linear or branching evolution patterns. Twenty-five patients proceeded to allogeneic stem cell transplantation. At a median follow-up of 21 months, median PFS and OS were not reached. In conclusion, patients who fail CPIs can be effectively rescued by salvage chemotherapy and bridged to allo-SCT/auto-SCT.

Large-scale analysis of SARS-CoV-2 synonymous mutations reveals the adaptation to the human codon usage during the virus evolution.

Many large national and transnational studies have been dedicated to the analysis of Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) genome, most of which focused on missense and nonsense mutations. However, approximately 30 per cent of the SARS-CoV-2 variants are synonymous, therefore changing the target codon without affecting the corresponding protein sequence. By performing a large-scale analysis of sequencing data generated from almost 400,000 SARS-CoV-2 samples, we show that silent mutations increasing the similarity of viral codons to the human ones tend to fixate in the viral genome overtime. This indicates that SARS-CoV-2 codon usage is adapting to the human host, likely improving its effectiveness in using the human aminoacyl-tRNA set through the accumulation of deceitfully neutral silent mutations. One-Sentence Summary. Synonymous SARS-CoV-2 mutations related to the activity of different mutational processes may positively impact viral evolution by increasing its adaptation to the human codon usage.

On the Use of Topological Features of Metabolic Networks for the Classification of Cancer Samples.

BACKGROUND: The increasing availability of omics data collected from patients affected by severe pathologies, such as cancer, is fostering the development of data science methods for their analysis. INTRODUCTION: The combination of data integration and machine learning approaches can provide new powerful instruments to tackle the complexity of cancer development and deliver effective diagnostic and prognostic strategies. METHODS: We explore the possibility of exploiting the topological properties of sample-specific metabolic networks as features in a supervised classification task. Such networks are obtained by projecting transcriptomic data from RNA-seq experiments on genome-wide metabolic models to define weighted networks modeling the overall metabolic activity of a given sample. RESULTS: We show the classification results on a labeled breast cancer dataset from the TCGA database, including 210 samples (cancer vs. normal). In particular, we investigate how the performance is affected by a threshold-based pruning of the networks by comparing Artificial Neural Networks, Support Vector Machines and Random Forests. Interestingly, the best classification performance is achieved within a small threshold range for all methods, suggesting that it might represent an effective choice to recover useful information while filtering out noise from data. Overall, the best accuracy is achieved with SVMs, which exhibit performances similar to those obtained when gene expression profiles are used as features. CONCLUSION: These findings demonstrate that the topological properties of sample-specific metabolic networks are effective in classifying cancer and normal samples, suggesting that useful information can be extracted from a relatively limited number of features.

VERSO: A comprehensive framework for the inference of robust phylogenies and the quantification of intra-host genomic diversity of viral samples.

We introduce VERSO, a two-step framework for the characterization of viral evolution from sequencing data of viral genomes, which is an improvement on phylogenomic approaches for consensus sequences. VERSO exploits an efficient algorithmic strategy to return robust phylogenies from clonal variant profiles, also in conditions of sampling limitations. It then leverages variant frequency patterns to characterize the intra-host genomic diversity of samples, revealing undetected infection chains and pinpointing variants likely involved in homoplasies. On simulations, VERSO outperforms state-of-the-art tools for phylogenetic inference. Notably, the application to 6,726 amplicon and RNA sequencing samples refines the estimation of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) evolution, while co-occurrence patterns of minor variants unveil undetected infection paths, which are validated with contact tracing data. Finally, the analysis of SARS-CoV-2 mutational landscape uncovers a temporal increase of overall genomic diversity and highlights variants transiting from minor to clonal state and homoplastic variants, some of which fall on the spike gene. Available at: https://github.com/BIMIB-DISCo/VERSO.

A review of computational strategies for denoising and imputation of single-cell transcriptomic data.

MOTIVATION: The advancements of single-cell sequencing methods have paved the way for the characterization of cellular states at unprecedented resolution, revolutionizing the investigation on complex biological systems. Yet, single-cell sequencing experiments are hindered by several technical issues, which cause output data to be noisy, impacting the reliability of downstream analyses. Therefore, a growing number of data science methods has been proposed to recover lost or corrupted information from single-cell sequencing data. To date, however, no quantitative benchmarks have been proposed to evaluate such methods. RESULTS: We present a comprehensive analysis of the state-of-the-art computational approaches for denoising and imputation of single-cell transcriptomic data, comparing their performance in different experimental scenarios. In detail, we compared 19 denoising and imputation methods, on both simulated and real-world datasets, with respect to several performance metrics related to imputation of dropout events, recovery of true expression profiles, characterization of cell similarity, identification of differentially expressed genes and computation time. The effectiveness and scalability of all methods were assessed with regard to distinct sequencing protocols, sample size and different levels of biological variability and technical noise. As a result, we identify a subset of versatile approaches exhibiting solid performances on most tests and show that certain algorithmic families prove effective on specific tasks but inefficient on others. Finally, most methods appear to benefit from the introduction of appropriate assumptions on noise distribution of biological processes.

An Optimal Control Framework for the Automated Design of Personalized Cancer Treatments.

One of the key challenges in current cancer research is the development of computational strategies to support clinicians in the identification of successful personalized treatments. Control theory might be an effective approach to this end, as proven by the long-established application to therapy design and testing. In this respect, we here introduce the Control Theory for Therapy Design (CT4TD) framework, which employs optimal control theory on patient-specific pharmacokinetics (PK) and pharmacodynamics (PD) models, to deliver optimized therapeutic strategies. The definition of personalized PK/PD models allows to explicitly consider the physiological heterogeneity of individuals and to adapt the therapy accordingly, as opposed to standard clinical practices. CT4TD can be used in two distinct scenarios. At the time of the diagnosis, CT4TD allows to set optimized personalized administration strategies, aimed at reaching selected target drug concentrations, while minimizing the costs in terms of toxicity and adverse effects. Moreover, if longitudinal data on patients under treatment are available, our approach allows to adjust the ongoing therapy, by relying on simplified models of cancer population dynamics, with the goal of minimizing or controlling the tumor burden. CT4TD is highly scalable, as it employs the efficient dCRAB/RedCRAB optimization algorithm, and the results are robust, as proven by extensive tests on synthetic data. Furthermore, the theoretical framework is general, and it might be applied to any therapy for which a PK/PD model can be estimated, and for any kind of administration and cost. As a proof of principle, we present the application of CT4TD to Imatinib administration in Chronic Myeloid leukemia, in which we adopt a simplified model of cancer population dynamics. In particular, we show that the optimized therapeutic strategies are diversified among patients, and display improvements with respect to the current standard regime.

MaREA4Galaxy: Metabolic reaction enrichment analysis and visualization of RNA-seq data within Galaxy.

We present MaREA4Galaxy, a user-friendly tool that allows a user to characterize and to graphically compare groups of samples with different transcriptional regulation of metabolism, as estimated from cross-sectional RNA-seq data. The tool is available as plug-in for the widely-used Galaxy platform for comparative genomics and bioinformatics analyses. MaREA4Galaxy combines three modules. The Expression2RAS module, which, for each reaction of a specified set, computes a Reaction Activity Score (RAS) as a function of the expression level of genes encoding for the associated enzyme. The MaREA (Metabolic Reaction Enrichment Analysis) module that allows to highlight significant differences in reaction activities between specified groups of samples. The Clustering module which employs the RAS computed before as a metric for unsupervised clustering of samples into distinct metabolic subgroups; the Clustering tool provides different clustering techniques and implements standard methods to evaluate the goodness of the results.

Learning mutational graphs of individual tumour evolution from single-cell and multi-region sequencing data.

BACKGROUND: A large number of algorithms is being developed to reconstruct evolutionary models of individual tumours from genome sequencing data. Most methods can analyze multiple samples collected either through bulk multi-region sequencing experiments or the sequencing of individual cancer cells. However, rarely the same method can support both data types. RESULTS: We introduce TRaIT, a computational framework to infer mutational graphs that model the accumulation of multiple types of somatic alterations driving tumour evolution. Compared to other tools, TRaIT supports multi-region and single-cell sequencing data within the same statistical framework, and delivers expressive models that capture many complex evolutionary phenomena. TRaIT improves accuracy, robustness to data-specific errors and computational complexity compared to competing methods. CONCLUSIONS: We show that the application of TRaIT to single-cell and multi-region cancer datasets can produce accurate and reliable models of single-tumour evolution, quantify the extent of intra-tumour heterogeneity and generate new testable experimental hypotheses.

Integration of transcriptomic data and metabolic networks in cancer samples reveals highly significant prognostic power.

Effective stratification of cancer patients on the basis of their molecular make-up is a key open challenge. Given the altered and heterogenous nature of cancer metabolism, we here propose to use the overall expression of central carbon metabolism as biomarker to characterize groups of patients with important characteristics, such as response to ad-hoc therapeutic strategies and survival expectancy. To this end, we here introduce the data integration framework named Metabolic Reaction Enrichment Analysis (MaREA), which strives to characterize the metabolic deregulations that distinguish cancer phenotypes, by projecting RNA-seq data onto metabolic networks, without requiring metabolic measurements. MaREA computes a score for each network reaction, based on the expression of the set of genes encoding for the associated enzyme(s). The scores are first used as features for cluster analysis and then to rank and visualize in an organized fashion the metabolic deregulations that distinguish cancer sub-types. We applied our method to recent lung and breast cancer RNA-seq datasets from The Cancer Genome Atlas and we were able to identify subgroups of patients with significant differences in survival expectancy. We show how the prognostic power of MaREA improves when an extracted and further curated core model focusing on central carbon metabolism is used rather than the genome-wide reference network. The visualization of the metabolic differences between the groups with best and worst prognosis allowed to identify and analyze key metabolic properties related to cancer aggressiveness. Some of these properties are shared across different cancer (sub) types, e.g., the up-regulation of nucleic acid and amino acid synthesis, whereas some other appear to be tumor-specific, such as the up- or down-regulation of the phosphoenolpyruvate carboxykinase reaction, which display different patterns in distinct tumor (sub)types. These results might be soon employed to deliver highly automated diagnostic and prognostic strategies for cancer patients.

Modeling Cumulative Biological Phenomena with Suppes-Bayes Causal Networks.

Several diseases related to cell proliferation are characterized by the accumulation of somatic DNA changes, with respect to wild-type conditions. Cancer and HIV are 2 common examples of such diseases, where the mutational load in the cancerous/viral population increases over time. In these cases, selective pressures are often observed along with competition, co-operation, and parasitism among distinct cellular clones. Recently, we presented a mathematical framework to model these phenomena, based on a combination of Bayesian inference and Suppes' theory of probabilistic causation, depicted in graphical structures dubbed Suppes-Bayes Causal Networks (SBCNs). The SBCNs are generative probabilistic graphical models that recapitulate the potential ordering of accumulation of such DNA changes during the progression of the disease. Such models can be inferred from data by exploiting likelihood-based model selection strategies with regularization. In this article, we discuss the theoretical foundations of our approach and we investigate in depth the influence on the model selection task of (1) the poset based on Suppes' theory and (2) different regularization strategies. Furthermore, we provide an example of application of our framework to HIV genetic data highlighting the valuable insights provided by the inferred SBCN.