A deep profile of gene expression across 18 human cancers.

Clinically and biologically valuable information may reside untapped in large cancer gene expression data sets. Deep unsupervised learning has the potential to extract this information with unprecedented efficacy but has thus far been hampered by a lack of biological interpretability and robustness. Here, we present DeepProfile, a comprehensive framework that addresses current challenges in applying unsupervised deep learning to gene expression profiles. We use DeepProfile to learn low-dimensional latent spaces for 18 human cancers from 50,211 transcriptomes. DeepProfile outperforms existing dimensionality reduction methods with respect to biological interpretability. Using DeepProfile interpretability methods, we show that genes that are universally important in defining the latent spaces across all cancer types control immune cell activation, while cancer type-specific genes and pathways define molecular disease subtypes. By linking DeepProfile latent variables to secondary tumor characteristics, we discover that tumor mutation burden is closely associated with the expression of cell cycle-related genes. DNA mismatch repair and MHC class II antigen presentation pathway expression, on the other hand, are consistently associated with patient survival. We validate these results through Kaplan-Meier analyses and nominate tumor-associated macrophages as an important source of survival-correlated MHC class II transcripts. Our results illustrate the power of unsupervised deep learning for discovery of novel cancer biology from existing gene expression data.

Uncovering expression signatures of synergistic drug responses via ensembles of explainable machine-learning models.

Machine learning may aid the choice of optimal combinations of anticancer drugs by explaining the molecular basis of their synergy. By combining accurate models with interpretable insights, explainable machine learning promises to accelerate data-driven cancer pharmacology. However, owing to the highly correlated and high-dimensional nature of transcriptomic data, naively applying current explainable machine-learning strategies to large transcriptomic datasets leads to suboptimal outcomes. Here by using feature attribution methods, we show that the quality of the explanations can be increased by leveraging ensembles of explainable machine-learning models. We applied the approach to a dataset of 133 combinations of 46 anticancer drugs tested in ex vivo tumour samples from 285 patients with acute myeloid leukaemia and uncovered a haematopoietic-differentiation signature underlying drug combinations with therapeutic synergy. Ensembles of machine-learning models trained to predict drug combination synergies on the basis of gene-expression data may improve the feature attribution quality of complex machine-learning models.

Reducing Peptide Sequence Bias in Quantitative Mass Spectrometry Data with Machine Learning.

Quantitative mass spectrometry measurements of peptides necessarily incorporate sequence-specific biases that reflect the behavior of the peptide during enzymatic digestion and liquid chromatography and in a mass spectrometer. These sequence-specific effects impair quantification accuracy, yielding peptide quantities that are systematically under- or overestimated. We provide empirical evidence for the existence of such biases, and we use a deep neural network, called Pepper, to automatically identify and reduce these biases. The model generalizes to new proteins and new runs within a related set of tandem mass spectrometry experiments, and the learned coefficients themselves reflect expected physicochemical properties of the corresponding peptide sequences. The resulting adjusted abundance measurements are more correlated with mRNA-based gene expression measurements than the unadjusted measurements. Pepper is suitable for data generated on a variety of mass spectrometry instruments and can be used with labeled or label-free approaches and with data-independent or data-dependent acquisition.

An automatic integrative method for learning interpretable communities of biological pathways.

Although knowledge of biological pathways is essential for interpreting results from computational biology studies, the growing number of pathway databases complicates efforts to efficiently perform pathway analysis due to high redundancies among pathways from different databases, and inconsistencies in how pathways are created and named. We introduce the PAthway Communities (PAC) framework, which reconciles pathways from different databases and reduces pathway redundancy by revealing informative groups with distinct biological functions. Uniquely applying the Louvain community detection algorithm to a network of 4847 pathways from KEGG, REACTOME and Gene Ontology databases, we identify 35 distinct and automatically annotated communities of pathways and show that they are consistent with expert-curated pathway categories. Further, we demonstrate that our pathway community network can be queried with new gene sets to provide biological context in terms of related pathways and communities. Our approach, combined with an interpretable web tool we provide, will help computational biologists more efficiently contextualize and interpret their biological findings.

Adversarial deconfounding autoencoder for learning robust gene expression embeddings.

MOTIVATION: Increasing number of gene expression profiles has enabled the use of complex models, such as deep unsupervised neural networks, to extract a latent space from these profiles. However, expression profiles, especially when collected in large numbers, inherently contain variations introduced by technical artifacts (e.g. batch effects) and uninteresting biological variables (e.g. age) in addition to the true signals of interest. These sources of variations, called confounders, produce embeddings that fail to transfer to different domains, i.e. an embedding learned from one dataset with a specific confounder distribution does not generalize to different distributions. To remedy this problem, we attempt to disentangle confounders from true signals to generate biologically informative embeddings. RESULTS: In this article, we introduce the Adversarial Deconfounding AutoEncoder (AD-AE) approach to deconfounding gene expression latent spaces. The AD-AE model consists of two neural networks: (i) an autoencoder to generate an embedding that can reconstruct original measurements, and (ii) an adversary trained to predict the confounder from that embedding. We jointly train the networks to generate embeddings that can encode as much information as possible without encoding any confounding signal. By applying AD-AE to two distinct gene expression datasets, we show that our model can (i) generate embeddings that do not encode confounder information, (ii) conserve the biological signals present in the original space and (iii) generalize successfully across different confounder domains. We demonstrate that AD-AE outperforms standard autoencoder and other deconfounding approaches. AVAILABILITY AND IMPLEMENTATION: Our code and data are available at https://gitlab.cs.washington.edu/abdincer/ad-ae. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.