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.

Chromatin Velocity reveals epigenetic dynamics by single-cell profiling of heterochromatin and euchromatin.

Recent efforts have succeeded in surveying open chromatin at the single-cell level, but high-throughput, single-cell assessment of heterochromatin and its underlying genomic determinants remains challenging. We engineered a hybrid transposase including the chromodomain (CD) of the heterochromatin protein-1α (HP-1α), which is involved in heterochromatin assembly and maintenance through its binding to trimethylation of the lysine 9 on histone 3 (H3K9me3), and developed a single-cell method, single-cell genome and epigenome by transposases sequencing (scGET-seq), that, unlike single-cell assay for transposase-accessible chromatin with sequencing (scATAC-seq), comprehensively probes both open and closed chromatin and concomitantly records the underlying genomic sequences. We tested scGET-seq in cancer-derived organoids and human-derived xenograft (PDX) models and identified genetic events and plasticity-driven mechanisms contributing to cancer drug resistance. Next, building upon the differential enrichment of closed and open chromatin, we devised a method, Chromatin Velocity, that identifies the trajectories of epigenetic modifications at the single-cell level. Chromatin Velocity uncovered paths of epigenetic reorganization during stem cell reprogramming and identified key transcription factors driving these developmental processes. scGET-seq reveals the dynamics of genomic and epigenetic landscapes underlying any cellular processes.

Nested Stochastic Block Models applied to the analysis of single cell data.

Single cell profiling has been proven to be a powerful tool in molecular biology to understand the complex behaviours of heterogeneous system. The definition of the properties of single cells is the primary endpoint of such analysis, cells are typically clustered to underpin the common determinants that can be used to describe functional properties of the cell mixture under investigation. Several approaches have been proposed to identify cell clusters; while this is matter of active research, one popular approach is based on community detection in neighbourhood graphs by optimisation of modularity. In this paper we propose an alternative and principled solution to this problem, based on Stochastic Block Models. We show that such approach not only is suitable for identification of cell groups, it also provides a solid framework to perform other relevant tasks in single cell analysis, such as label transfer. To encourage the use of Stochastic Block Models, we developed a python library, schist, that is compatible with the popular scanpy framework.

Fast analysis of scATAC-seq data using a predefined set of genomic regions.

Background: Analysis of scATAC-seq data has been recently scaled to thousands of cells. While processing of other types of single cell data was boosted by the implementation of alignment-free techniques, pipelines available to process scATAC-seq data still require large computational resources. We propose here an approach based on pseudoalignment, which reduces the execution times and hardware needs at little cost for precision. Methods: Public data for 10k PBMC were downloaded from 10x Genomics web site. Reads were aligned to various references derived from DNase I Hypersensitive Sites (DHS) using kallisto and quantified with bustools. We compared our results with the ones publicly available derived by cellranger-atac. Results: We found that kallisto does not introduce biases in quantification of known peaks and cells groups are identified in a consistent way. We also found that cell identification is robust when analysis is performed using DHS-derived reference in place of de novo identification of ATAC peaks. Lastly, we found that our approach is suitable for reliable quantification of gene activity based on scATAC-seq signal, thus allows for efficient labelling of cell groups based on marker genes. Conclusions: Analysis of scATAC-seq data by means of kallisto produces results in line with standard pipelines while being considerably faster; using a set of known DHS sites as reference does not affect the ability to characterize the cell populations.

Integration of Machine Learning Methods to Dissect Genetically Imputed Transcriptomic Profiles in Alzheimer's Disease.

The genetic component of many common traits is associated with the gene expression and several variants act as expression quantitative loci, regulating the gene expression in a tissue specific manner. In this work, we applied tissue-specific cis-eQTL gene expression prediction models on the genotype of 808 samples including controls, subjects with mild cognitive impairment, and patients with Alzheimer's Disease. We then dissected the imputed transcriptomic profiles by means of different unsupervised and supervised machine learning approaches to identify potential biological associations. Our analysis suggests that unsupervised and supervised methods can provide complementary information, which can be integrated for a better characterization of the underlying biological system. In particular, a variational autoencoder representation of the transcriptomic profiles, followed by a support vector machine classification, has been used for tissue-specific gene prioritizations. Interestingly, the achieved gene prioritizations can be efficiently integrated as a feature selection step for improving the accuracy of deep learning classifier networks. The identified gene-tissue information suggests a potential role for inflammatory and regulatory processes in gut-brain axis related tissues. In line with the expected low heritability that can be apportioned to eQTL variants, we were able to achieve only relatively low prediction capability with deep learning classification models. However, our analysis revealed that the classification power strongly depends on the network structure, with recurrent neural networks being the best performing network class. Interestingly, cross-tissue analysis suggests a potentially greater role of models trained in brain tissues also by considering dementia-related endophenotypes. Overall, the present analysis suggests that the combination of supervised and unsupervised machine learning techniques can be used for the evaluation of high dimensional omics data.

Improving eQTL Analysis Using a Machine Learning Approach for Data Integration: A Logistic Model Tree Solution.

Expression quantitative trait loci (eQTL) analysis is an emerging method for establishing the impact of genetic variations (such as single nucleotide polymorphisms) on the expression levels of genes. Although different methods for evaluating the impact of these variations are proposed in the literature, the results obtained are mostly in disagreement, entailing a considerable number of false-positive predictions. For this reason, we propose an approach based on Logistic Model Trees that integrates the predictions of different eQTL mapping tools to produce more reliable results. More precisely, we employ a machine learning-based method using logistic functions to perform a linear regression able to classify the predictions of three eQTL analysis tools (namely, R/qtl, MatrixEQTL, and mRMR). Given the lack of a reference dataset and that computational predictions are not so easy to test experimentally, the performance of our approach is assessed using data from the DREAM5 challenge. The results show the quality of the aggregated prediction is better than that obtained by each single tool in terms of both precision and recall. We also performed a test on real data, employing genotypes and microRNA expression profiles from Caenorhabditis elegans, which proved that we were able to correctly classify all the experimentally validated eQTLs. These good results come both from the integration of the different predictions, and from the ability of this machine learning algorithm to find the best cutoff thresholds for each tool. This combination makes our integration approach suitable for improving eQTL predictions for testing in a laboratory, reducing the number of false-positive results.