A hybrid deep clustering approach for robust cell type profiling using single-cell RNA-seq data.

Single-cell RNA sequencing (scRNA-seq) is a recent technology that enables fine-grained discovery of cellular subtypes and specific cell states. Analysis of scRNA-seq data routinely involves machine learning methods, such as feature learning, clustering, and classification, to assist in uncovering novel information from scRNA-seq data. However, current methods are not well suited to deal with the substantial amount of noise that is created by the experiments or the variation that occurs due to differences in the cells of the same type. To address this, we developed a new hybrid approach, deep unsupervised single-cell clustering (DUSC), which integrates feature generation based on a deep learning architecture by using a new technique to estimate the number of latent features, with a model-based clustering algorithm, to find a compact and informative representation of the single-cell transcriptomic data generating robust clusters. We also include a technique to estimate an efficient number of latent features in the deep learning model. Our method outperforms both classical and state-of-the-art feature learning and clustering methods, approaching the accuracy of supervised learning. We applied DUSC to a single-cell transcriptomics data set obtained from a triple-negative breast cancer tumor to identify potential cancer subclones accentuated by copy-number variation and investigate the role of clonal heterogeneity. Our method is freely available to the community and will hopefully facilitate our understanding of the cellular atlas of living organisms as well as provide the means to improve patient diagnostics and treatment.

Biological classification with RNA-seq data: Can alternatively spliced transcript expression enhance machine learning classifiers?

RNA sequencing (RNA-seq) is becoming a prevalent approach to quantify gene expression and is expected to gain better insights into a number of biological and biomedical questions compared to DNA microarrays. Most importantly, RNA-seq allows us to quantify expression at the gene or transcript levels. However, leveraging the RNA-seq data requires development of new data mining and analytics methods. Supervised learning methods are commonly used approaches for biological data analysis that have recently gained attention for their applications to RNA-seq data. Here, we assess the utility of supervised learning methods trained on RNA-seq data for a diverse range of biological classification tasks. We hypothesize that the transcript-level expression data are more informative for biological classification tasks than the gene-level expression data. Our large-scale assessment utilizes multiple data sets, organisms, lab groups, and RNA-seq analysis pipelines. Overall, we performed and assessed 61 biological classification problems that leverage three independent RNA-seq data sets and include over 2000 samples that come from multiple organisms, lab groups, and RNA-seq analyses. These 61 problems include predictions of the tissue type, sex, or age of the sample, healthy or cancerous phenotypes, and pathological tumor stages for the samples from the cancerous tissue. For each problem, the performance of three normalization techniques and six machine learning classifiers was explored. We find that for every single classification problem, the transcript-based classifiers outperform or are comparable with gene expression-based methods. The top-performing techniques reached a near perfect classification accuracy, demonstrating the utility of supervised learning for RNA-seq based data analysis.

The genome of the soybean cyst nematode (Heterodera glycines) reveals complex patterns of duplications involved in the evolution of parasitism genes.

BACKGROUND: Heterodera glycines, commonly referred to as the soybean cyst nematode (SCN), is an obligatory and sedentary plant parasite that causes over a billion-dollar yield loss to soybean production annually. Although there are genetic determinants that render soybean plants resistant to certain nematode genotypes, resistant soybean cultivars are increasingly ineffective because their multi-year usage has selected for virulent H. glycines populations. The parasitic success of H. glycines relies on the comprehensive re-engineering of an infection site into a syncytium, as well as the long-term suppression of host defense to ensure syncytial viability. At the forefront of these complex molecular interactions are effectors, the proteins secreted by H. glycines into host root tissues. The mechanisms of effector acquisition, diversification, and selection need to be understood before effective control strategies can be developed, but the lack of an annotated genome has been a major roadblock. RESULTS: Here, we use PacBio long-read technology to assemble a H. glycines genome of 738 contigs into 123 Mb with annotations for 29,769 genes. The genome contains significant numbers of repeats (34%), tandem duplicates (18.7 Mb), and horizontal gene transfer events (151 genes). A large number of putative effectors (431 genes) were identified in the genome, many of which were found in transposons. CONCLUSIONS: This advance provides a glimpse into the host and parasite interplay by revealing a diversity of mechanisms that give rise to virulence genes in the soybean cyst nematode, including: tandem duplications containing over a fifth of the total gene count, virulence genes hitchhiking in transposons, and 107 horizontal gene transfers not reported in other plant parasitic nematodes thus far. Through extensive characterization of the H. glycines genome, we provide new insights into H. glycines biology and shed light onto the mystery underlying complex host-parasite interactions. This genome sequence is an important prerequisite to enable work towards generating new resistance or control measures against H. glycines.

Conditional deletion of calcium-modulating cyclophilin ligand causes deafness in mice.

Calcium-modulating cyclophilin ligand (Caml) is a ubiquitously expressed cytoplasmic protein that is involved in multiple signaling and developmental pathways. An observation in our laboratory of a protein-protein interaction between Caml and the cytoplasmic region of Cadherin23 led us to speculate that Caml might be important in the inner ear and play a role in the development and/or function of hair cells. To address this question, we generated a mouse line in which Caml expression was eliminated in Atoh1-expressing cells of the inner ear upon administration of tamoxifen. Tamoxifen was administered immediately after birth to neonates to assess the effect of loss of Caml in the inner ear during postnatal development. Hearing in treated animals was tested by auditory brain stem response (ABR) analysis and cochlear pathology was evaluated by light microscopy. Lack of Caml expression in the inner ear leads to severe loss of cochlear hair cells and complete deafness. Elucidating the role of Caml in the inner ear will aid our understanding of the molecular pathways important for auditory development and function.

Predicting protein interaction network perturbation by alternative splicing with semi-supervised learning.

Alternative splicing introduces an additional layer of protein diversity and complexity in regulating cellular functions that can be specific to the tissue and cell type, physiological state of a cell, or disease phenotype. Recent high-throughput experimental studies have illuminated the functional role of splicing events through rewiring protein-protein interactions; however, the extent to which the macromolecular interactions are affected by alternative splicing has yet to be fully understood. In silico methods provide a fast and cheap alternative to interrogating functional characteristics of thousands of alternatively spliced isoforms. Here, we develop an accurate feature-based machine learning approach that predicts whether a protein-protein interaction carried out by a reference isoform is perturbed by an alternatively spliced isoform. Our method, called the alternatively spliced interactions prediction (ALT-IN) tool, is compared with the state-of-the-art PPI prediction tools and shows superior performance, achieving 0.92 in precision and recall values.

Machine learning identifies candidates for drug repurposing in Alzheimer's disease.

Clinical trials of novel therapeutics for Alzheimer's Disease (AD) have consumed a large amount of time and resources with largely negative results. Repurposing drugs already approved by the Food and Drug Administration (FDA) for another indication is a more rapid and less expensive option. We present DRIAD (Drug Repurposing In AD), a machine learning framework that quantifies potential associations between the pathology of AD severity (the Braak stage) and molecular mechanisms as encoded in lists of gene names. DRIAD is applied to lists of genes arising from perturbations in differentiated human neural cell cultures by 80 FDA-approved and clinically tested drugs, producing a ranked list of possible repurposing candidates. Top-scoring drugs are inspected for common trends among their targets. We propose that the DRIAD method can be used to nominate drugs that, after additional validation and identification of relevant pharmacodynamic biomarker(s), could be readily evaluated in a clinical trial.

Targeting immunosuppressive macrophages overcomes PARP inhibitor resistance in BRCA1-associated triple-negative breast cancer.

Despite objective responses to PARP inhibition and improvements in progression-free survival compared to standard chemotherapy in patients with BRCA-associated triple-negative breast cancer (TNBC), benefits are transitory. Using high dimensional single-cell profiling of human TNBC, here we demonstrate that macrophages are the predominant infiltrating immune cell type in BRCA-associated TNBC. Through multi-omics profiling we show that PARP inhibitors enhance both anti- and pro-tumor features of macrophages through glucose and lipid metabolic reprogramming driven by the sterol regulatory element-binding protein 1 (SREBP-1) pathway. Combined PARP inhibitor therapy with CSF-1R blocking antibodies significantly enhanced innate and adaptive anti-tumor immunity and extends survival in BRCA-deficient tumors in vivo and is mediated by CD8+ T-cells. Collectively, our results uncover macrophage-mediated immune suppression as a liability of PARP inhibitor treatment and demonstrate combined PARP inhibition and macrophage targeting therapy induces a durable reprogramming of the tumor microenvironment, thus constituting a promising therapeutic strategy for TNBC.

Inferring a role for methylation of intergenic DNA in the regulation of genes aberrantly expressed in precursor B-cell acute lymphoblastic leukemia.

A complete understanding of the mechanisms involved in the development of pre-B ALL is lacking. In this study, we integrated DNA methylation data and gene expression data to elucidate the impact of aberrant intergenic DNA methylation on gene expression in pre-B ALL. We found a subset of differentially methylated intergenic loci that were associated with altered gene expression in pre-B ALL patients. Notably, 84% of these regions were also bound by transcription factors (TF) known to play roles in differentiation and B-cell development in a lymphoblastoid cell line. Further, an overall downregulation of eRNA transcripts was observed in pre-B ALL patients and these transcripts were associated with the downregulation of putative target genes involved in B-cell migration, proliferation, and apoptosis. The identification of novel putative regulatory regions highlights the significance of intergenic DNA sequences and may contribute to the identification of new therapeutic targets for the treatment of pre-B ALL.

Integrated methylome and transcriptome analysis reveals novel regulatory elements in pediatric acute lymphoblastic leukemia.

Acute lymphoblastic leukemia (ALL) is the most common cancer diagnosed in children under the age of 15. In addition to genetic aberrations, epigenetic modifications such as DNA methylation are altered in cancer and impact gene expression. To identify epigenetic alterations in ALL, genome-wide methylation profiles were generated using the methylated CpG island recovery assay followed by next-generation sequencing. More than 25,000 differentially methylated regions (DMR) were observed in ALL patients with ∼ 90% present within intronic or intergenic regions. To determine the regulatory potential of the DMR, whole-transcriptome analysis was performed and integrated with methylation data. Aberrant promoter methylation was associated with the altered expression of genes involved in transcriptional regulation, apoptosis, and proliferation. Novel enhancer-like sequences were identified within intronic and intergenic DMR. Aberrant methylation in these regions was associated with the altered expression of neighboring genes involved in cell cycle processes, lymphocyte activation and apoptosis. These genes include potential epi-driver genes, such as SYNE1, PTPRS, PAWR, HDAC9, RGCC, MCOLN2, LYN, TRAF3, FLT1, and MELK, which may provide a selective advantage to leukemic cells. In addition, the differential expression of epigenetic modifier genes, pseudogenes, and non-coding RNAs was also observed accentuating the role of erroneous epigenetic gene regulation in ALL.