The epigenome involves a complex set of cellular processes governing genomic activity. Dissecting this complexity necessitates the development of tools capable of specifically manipulating these processes. The repurposing of prokaryotic CRISPR systems has allowed for the development of diverse technologies for epigenome engineering. Here, we review the state of currently achievable epigenetic manipulations along with corresponding applications. With future optimization, CRISPR-based epigenomic editing stands as a set of powerful tools for understanding and controlling biological function. <<<

Jae-Hyun Yang, Motoshi Hayano, Patrick T. Griffin, João A. Amorim, Michael S. Bonkowski, John K. Apostolides, Elias L. Salfati, Marco Blanchette, Elizabeth M. Munding, Mital Bhakta, Yap Ching Chew, Wei Guo, Xiaojing Yang, Sun Maybury-Lewis, Xiao Tian, Jaime M. Ross, Giuseppe Coppotelli, Margarita V. Meer, Ryan Rogers-Hammond, Daniel L. Vera, Yuancheng Ryan Lu, Jeffrey W. Pippin, Michael L. Creswell, Zhixun Dou, Caiyue Xu, Sarah J. Mitchell, Abhirup Das, Brendan L. O’Connell, Sachin Thakur, Alice E. Kane, Qiao Su, Yasuaki Mohri, Emi K. Nishimura, Laura Schaevitz, Neha Garg, Ana-Maria Balta, Meghan A. Rego, Meredith Gregory-Ksander, Tatjana C. Jakobs, Lei Zhong, Hiroko Wakimoto, Jihad El Andari, Dirk Grimm, Raul Mostoslavsky, Amy J. Wagers, Kazuo Tsubota, Stephen J. Bonasera, Carlos M. Palmeira, Jonathan G. Seidman, Christine E. Seidman, Norman S. Wolf, Jill A. Kreiling, John M. Sedivy, George F. Murphy, Richard E. Green, Benjamin A. Garcia, Shelley L. Berger, Philipp Oberdoerffer, Stuart J. Shankland, Vadim N. Gladyshev, Bruce R. Ksander, Andreas R. Pfenning, Luis A. Rajman, David A. Sinclair <<<

BACKGROUND: There is divergence in the rate at which people age. The concept of biological age is postulated to capture this variability, and hence to better represent an individual's true global physiological state than chronological age. Biological age predictors are often generated based on cross-sectional data, using biochemical or molecular markers as predictor variables. It is assumed that the difference between chronological and predicted biological age is informative of one's chronological age-independent aging divergence ∆.METHODS: We investigated the statistical assumptions underlying the most popular cross-sectional biological age predictors, based on multiple linear regression, the Klemera-Doubal method or principal component analysis. We used synthetic and real data to illustrate the consequences if this assumption does not hold.RESULTS: The most popular cross-sectional biological age predictors all use the same strong underlying assumption, namely that a candidate marker of aging's association with chronological age is directly informative of its association with the aging rate ∆. We called this the identical-association assumption and proved that it is untestable in a cross-sectional setting. If this assumption does not hold, weights assigned to candidate markers of aging are uninformative, and no more signal may be captured than if markers would have been assigned weights at random.CONCLUSIONS: Cross-sectional methods for predicting biological age commonly use the untestable identical-association assumption, which previous literature in the field had never explicitly acknowledged. These methods have inherent limitations and may provide uninformative results, highlighting the importance of researchers exercising caution in the development and interpretation of cross-sectional biological age predictors. <<<

Christina V Theodoris, Ling Xiao, Anant Chopra, Mark D Chaffin, Zeina R Al Sayed, Matthew C Hill, Helene Mantineo, Elizabeth M Brydon, Zexian Zeng, X Shirley Liu, Patrick T Ellinor <<<

Mapping gene networks requires large amounts of transcriptomic data to learn the connections between genes, which impedes discoveries in settings with limited data, including rare diseases and diseases affecting clinically inaccessible tissues. Recently, transfer learning has revolutionized fields such as natural language understanding and computer vision by leveraging deep learning models pretrained on large-scale general datasets that can then be fine-tuned towards a vast array of downstream tasks with limited task-specific data. Here, we developed a context-aware, attention-based deep learning model, Geneformer, pretrained on a large-scale corpus of about 30 million single-cell transcriptomes to enable context-specific predictions in settings with limited data in network biology. During pretraining, Geneformer gained a fundamental understanding of network dynamics, encoding network hierarchy in the attention weights of the model in a completely self-supervised manner. Fine-tuning towards a diverse panel of downstream tasks relevant to chromatin and network dynamics using limited task-specific data demonstrated that Geneformer consistently boosted predictive accuracy. Applied to disease modelling with limited patient data, Geneformer identified candidate therapeutic targets for cardiomyopathy. Overall, Geneformer represents a pretrained deep learning model from which fine-tuning towards a broad range of downstream applications can be pursued to accelerate discovery of key network regulators and candidate therapeutic targets. <<<

Prediction-powered inference is a framework for performing valid statistical inference when an experimental dataset is supplemented with predictions from a machine-learning system. The framework yields simple algorithms for computing provably valid confidence intervals for quantities such as means, quantiles, and linear and logistic regression coefficients without making any assumptions about the machine-learning algorithm that supplies the predictions. Furthermore, more accurate predictions translate to smaller confidence intervals. Prediction-powered inference could enable researchers to draw valid and more data-efficient conclusions using machine learning. The benefits of prediction-powered inference were demonstrated with datasets from proteomics, astronomy, genomics, remote sensing, census analysis, and ecology. <<<

Variational autoencoders are powerful algorithms for identifying dominantlatent structure in a single dataset. In many applications, however, we areinterested in modeling latent structure and variation that are enriched in atarget dataset compared to some background---e.g. enriched in patients comparedto the general population. Contrastive learning is a principled framework tocapture such enriched variation between the target and background, butstate-of-the-art contrastive methods are limited to linear models. In thispaper, we introduce the contrastive variational autoencoder (cVAE), whichcombines the benefits of contrastive learning with the power of deep generativemodels. The cVAE is designed to identify and enhance salient latent features.The cVAE is trained on two related but unpaired datasets, one of which hasminimal contribution from the salient latent features. The cVAE explicitlymodels latent features that are shared between the datasets, as well as thosethat are enriched in one dataset relative to the other, which allows thealgorithm to isolate and enhance the salient latent features. The algorithm isstraightforward to implement, has a similar run-time to the standard VAE, andis robust to noise and dataset purity. We conduct experiments across diversetypes of data, including gene expression and facial images, showing that thecVAE effectively uncovers latent structure that is salient in a particularanalysis. <<<

Over time, the human DNA methylation landscape accrues substantial damage, which has been associated with a broad range of age-related diseases, including cardiovascular disease and cancer. Various age-related DNA methylation changes have been described, including at the level of individual CpGs, such as differential and variable methylation, and at the level of the whole methylome, including entropy and correlation networks. Here, we review these changes in the ageing methylome as well as the statistical tools that can be used to quantify them. We detail the evidence linking DNA methylation to ageing phenotypes and the longevity strategies aimed at altering both DNA methylation patterns and machinery to extend healthspan and lifespan. Lastly, we discuss theories on the mechanistic causes of epigenetic ageing. <<<

An outstanding challenge of epigenome-wide association studies (EWASs) performed in complex tissues is the identification of the specific cell type(s) responsible for the observed differential DNA methylation. Here we present a statistical algorithm called CellDMC ( https://github.com/sjczheng/EpiDISH ), which can identify differentially methylated positions and the specific cell type(s) driving the differential methylation. We validated CellDMC on in silico mixtures of DNA methylation data generated with different technologies, as well as on real mixtures from epigenome-wide association and cancer epigenome studies. CellDMC achieved over 90% sensitivity and specificity in scenarios where current state-of-the-art methods did not identify differential methylation. By applying CellDMC to an EWAS performed in buccal swabs, we identified smoking-associated differentially methylated positions occurring in the epithelial compartment, which we validated in smoking-related lung cancer. CellDMC may be useful in the identification of causal DNA-methylation alterations in disease. <<<

Dynamic feature selection, where we sequentially query features to make accurate predictions with a minimal budget, is a promising paradigm to reduce feature acquisition costs and provide transparency into the prediction process. The problem is challenging, however, as it requires both making predictions with arbitrary feature sets and learning a policy to identify the most valuable selections. Here, we take an information-theoretic perspective and prioritize features based on their mutual information with the response variable. The main challenge is learning this selection policy, and we design a straightforward new modeling approach that estimates the mutual information in a discriminative rather than generative fashion. Building on our learning approach, we introduce several further improvements: allowing variable feature budgets across samples, enabling non-uniform costs between features, incorporating prior information, and exploring modern architectures to handle partial input information. We find that our method provides consistent gains over recent state-of-the-art methods across a variety of datasets. <<<

Tianshi Lu, Ze Zhang, James Zhu, Yunguan Wang, Peixin Jiang, Xue Xiao, Chantale Bernatchez, John V Heymach, Don L Gibbons, Jun Wang, Lin Xu, Alexandre Reuben, Tao Wang <<<

Neoantigens play a key role in the recognition of tumor cells by T cells. However, only a small proportion of neoantigens truly elicit T cell responses, and fewer clues exist as to which neoantigens are recognized by which T cell receptors (TCRs). We built a transfer learning-based model, named pMHC-TCR binding prediction network (pMTnet), to predict TCR-binding specificities of neoantigens, and T cell antigens in general, presented by class I major histocompatibility complexes (pMHCs). pMTnet was comprehensively validated by a series of analyses, and showed advance over previous work by a large margin. By applying pMTnet in human tumor genomics data, we discovered that neoantigens were generally more immunogenic than self-antigens, but HERV-E, a special type of self-antigen that is re-activated in kidney cancer, is more immunogenic than neoantigens. We further discovered that patients with more clonally expanded T cells exhibiting better affinity against truncal, rather than subclonal, neoantigens, had more favorable prognosis and treatment response to immunotherapy, in melanoma and lung cancer but not in kidney cancer. Predicting TCR-neoantigen/antigen pairs is one of the most daunting challenges in modern immunology. However, we achieved an accurate prediction of the pairing only using the TCR sequence (CDR3β), antigen sequence, and class I MHC allele, and our work revealed unique insights into the interactions of TCRs and pMHCs in human tumors using pMTnet as a discovery tool. <<<

Visualization and exploration of high-dimensional data is a ubiquitous challenge across disciplines. Widely used techniques such as principal component analysis (PCA) aim to identify dominant trends in one dataset. However, in many settings we have datasets collected under different conditions, e.g., a treatment and a control experiment, and we are interested in visualizing and exploring patterns that are specific to one dataset. This paper proposes a method, contrastive principal component analysis (cPCA), which identifies low-dimensional structures that are enriched in a dataset relative to comparison data. In a wide variety of experiments, we demonstrate that cPCA with a background dataset enables us to visualize dataset-specific patterns missed by PCA and other standard methods. We further provide a geometric interpretation of cPCA and strong mathematical guarantees. An implementation of cPCA is publicly available, and can be used for exploratory data analysis in many applications where PCA is currently used. <<<

SummaryVariable selection plays an important role in high dimensional statistical modelling which nowadays appears in many areas and is key to various scientific discoveries. For problems of large scale or dimensionality p, accuracy of estimation and computational cost are two top concerns. Recently, Candes and Tao have proposed the Dantzig selector using L1-regularization and showed that it achieves the ideal risk up to a logarithmic factor log(p). Their innovative procedure and remarkable result are challenged when the dimensionality is ultrahigh as the factor log(p) can be large and their uniform uncertainty principle can fail. Motivated by these concerns, we introduce the concept of sure screening and propose a sure screening method that is based on correlation learning, called sure independence screening, to reduce dimensionality from high to a moderate scale that is below the sample size. In a fairly general asymptotic framework, correlation learning is shown to have the sure screening property for even exponentially growing dimensionality. As a methodological extension, iterative sure independence screening is also proposed to enhance its finite sample performance. With dimension reduced accurately from high to below sample size, variable selection can be improved on both speed and accuracy, and can then be accomplished by a well-developed method such as smoothly clipped absolute deviation, the Dantzig selector, lasso or adaptive lasso. The connections between these penalized least squares methods are also elucidated. <<<

Lukas Heumos, Anna C Schaar, Christopher Lance, Anastasia Litinetskaya, Felix Drost, Luke Zappia, Malte D Lücken, Daniel C Strobl, Juan Henao, Fabiola Curion, Single-cell Best Practices Consortium, Herbert B Schiller, Fabian J Theis <<<

Recent advances in single-cell technologies have enabled high-throughput molecular profiling of cells across modalities and locations. Single-cell transcriptomics data can now be complemented by chromatin accessibility, surface protein expression, adaptive immune receptor repertoire profiling and spatial information. The increasing availability of single-cell data across modalities has motivated the development of novel computational methods to help analysts derive biological insights. As the field grows, it becomes increasingly difficult to navigate the vast landscape of tools and analysis steps. Here, we summarize independent benchmarking studies of unimodal and multimodal single-cell analysis across modalities to suggest comprehensive best-practice workflows for the most common analysis steps. Where independent benchmarks are not available, we review and contrast popular methods. Our article serves as an entry point for novices in the field of single-cell (multi-)omic analysis and guides advanced users to the most recent best practices. <<<

Learning a faithful directed acyclic graph (DAG) from samples of a joint distribution is a challenging combinatorial problem, owing to the intractable search space superexponential in the number of graph nodes. A recent breakthrough formulates the problem as a continuous optimization with a structural constraint that ensures acyclicity (Zheng et al., 2018). The authors apply the approach to the linear structural equation model (SEM) and the least-squares loss function that are statistically well justified but nevertheless limited. Motivated by the widespread success of deep learning that is capable of capturing complex nonlinear mappings, in this work we propose a deep generative model and apply a variant of the structural constraint to learn the DAG. At the heart of the generative model is a variational autoencoder parameterized by a novel graph neural network architecture, which we coin DAG-GNN. In addition to the richer capacity, an advantage of the proposed model is that it naturally handles discrete variables as well as vector-valued ones. We demonstrate that on synthetic data sets, the proposed method learns more accurate graphs for nonlinearly generated samples; and on benchmark data sets with discrete variables, the learned graphs are reasonably close to the global optima. The code is available at \url{this https URL}. <<<

DNA methylation is one of the most commonly studied epigenetic marks, due to its role in disease and development. Illumina methylation arrays have been extensively used to measure methylation across the human genome. Methylation array analysis has primarily focused on preprocessing, normalization, and identification of differentially methylated CpGs and regions. GOmeth and GOregion are new methods for performing unbiased gene set testing following differential methylation analysis. Benchmarking analyses demonstrate GOmeth outperforms other approaches, and GOregion is the first method for gene set testing of differentially methylated regions. Both methods are publicly available in the missMethyl Bioconductor R package. <<<

Christopher G Bell, Robert Lowe, Peter D Adams, Andrea A Baccarelli, Stephan Beck, Jordana T Bell, Brock C Christensen, Vadim N Gladyshev, Bastiaan T Heijmans, Steve Horvath, Trey Ideker, Jean-Pierre J Issa, Karl T Kelsey, Riccardo E Marioni, Wolf Reik, Caroline L Relton, Leonard C Schalkwyk, Andrew E Teschendorff, Wolfgang Wagner, Kang Zhang, Vardhman K Rakyan <<<

Epigenetic clocks comprise a set of CpG sites whose DNA methylation levels measure subject age. These clocks are acknowledged as a highly accurate molecular correlate of chronological age in humans and other vertebrates. Also, extensive research is aimed at their potential to quantify biological aging rates and test longevity or rejuvenating interventions. Here, we discuss key challenges to understand clock mechanisms and biomarker utility. This requires dissecting the drivers and regulators of age-related changes in single-cell, tissue- and disease-specific models, as well as exploring other epigenomic marks, longitudinal and diverse population studies, and non-human models. We also highlight important ethical issues in forensic age determination and predicting the trajectory of biological aging in an individual. <<<

Differential expression (DE) analysis and gene set enrichment (GSE) analysis are commonly applied in single cell RNA sequencing (scRNA-seq) studies. Here, we develop an integrative and scalable computational method, iDEA, to perform joint DE and GSE analysis through a hierarchical Bayesian framework. By integrating DE and GSE analyses, iDEA can improve the power and consistency of DE analysis and the accuracy of GSE analysis. Importantly, iDEA uses only DE summary statistics as input, enabling effective data modeling through complementing and pairing with various existing DE methods. We illustrate the benefits of iDEA with extensive simulations. We also apply iDEA to analyze three scRNA-seq data sets, where iDEA achieves up to five-fold power gain over existing GSE methods and up to 64% power gain over existing DE methods. The power gain brought by iDEA allows us to identify many pathways that would not be identified by existing approaches in these data. <<<