We consider the problem of causal discovery from longitudinal observational data. We develop a novel framework that simultaneously discovers the time-lagged causality and the possibly cyclic instantaneous causality. Under common causal discovery assumptions, combined with additional instrumental information typically available in longitudinal data, we prove the proposed model is generally identifiable. To the best of our knowledge, this is the first causal identification theory for directed graphs with general cyclic patterns that achieves unique causal identifiability. Structural learning is carried out in a fully Bayesian fashion. Through extensive simulations and an application to the Women's Interagency HIV Study, we demonstrate the identifiability, utility, and superiority of the proposed model against state-of-the-art alternative methods.

We study the sparse high-dimensional Gaussian mixture model when the number of clusters is allowed to grow with the sample size. A minimax lower bound for parameter estimation is established, and we show that a constrained maximum likelihood estimator achieves the minimax lower bound. However, this optimization-based estimator is computationally intractable because the objective function is highly nonconvex and the feasible set involves discrete structures. To address the computational challenge, we propose a computationally tractable Bayesian approach to estimate high-dimensional Gaussian mixtures whose cluster centers exhibit sparsity using a continuous spike-and-slab prior. We further prove that the posterior contraction rate of the proposed Bayesian method is minimax optimal. The mis- clustering rate is obtained as a by-product using tools from matrix perturbation theory. The proposed Bayesian sparse Gaussian mixture model does not require pre-specifying the number of clusters, which can be adaptively estimated. The validity and usefulness of the proposed method is demonstrated through simulation studies and the analysis of a real-world single-cell RNA sequencing data set.

In this paper, we propose a policy gradient method for confounded partially observable Markov decision processes (POMDPs) with continuous state and observation spaces in the offline setting. We first establish a novel identification result to non-parametrically estimate any history-dependent policy gradient under POMDPs using the offline data. The identification enables us to solve a sequence of conditional moment restrictions and adopt the min-max learning procedure with general function approximation for estimating the policy gradient. We then provide a finite-sample non-asymptotic bound for estimating the gradient uniformly over a pre-specified policy class in terms of the sample size, length of horizon, concentratability coefficient and the measure of ill-posedness in solving the conditional moment restrictions. Lastly, by deploying the proposed gradient estimation in the gradient ascent algorithm, we show the global convergence of the proposed algorithm in finding the history-dependent optimal policy under some technical conditions. To the best of our knowledge, this is the first work studying the policy gradient method for POMDPs under the offline setting.

We propose a model-based offline reinforcement learning (RL) algorithm for confounded partially observable Markov decision processes (POMDPs) under general function approximations and show it is provably efficient under some technical conditions such as the partial coverage imposed on the offline data distribution. Specifically, we first establish a novel model-based identification result for learning the effect of any action on the reward and future transitions in the confounded POMDP. Using this identification result, we then design a nonparametric two-stage estimation procedure to construct an estimator for off-policy evaluation (OPE), which permits general function approximations. Finally, we learn the optimal policy by performing a conservative policy optimization within the confidence regions based on the proposed estimation procedure for OPE. Under some mild conditions, we establish a finite-sample upper bound on the suboptimality of the learned policy in finding the optimal one, which depends on the sample size and the length of horizons polynomially.

Offline reinforcement learning (RL) is a promising approach for training intelligent medical agents to learn treatment policies and assist decision making in many healthcare applications, such as scheduling clinical visits and assigning dosages for patients with chronic conditions. In this paper, we investigate the potential usefulness of Decision Transformer (Chen et al., 2021)–a new offline RL paradign– in medical domains where decision making in continuous time is desired. As Decision Transformer only handles discrete-time (or turn-based) sequential decision making scenarios, we generalize it to Continuous-Time Decision Transformer that not only considers the past clinical measurements and treatments but also the timings of previous visits, and learns to suggest the timings of future visits as well as the treatment plan at each visit. Extensive experiments on synthetic datasets and simulators motivated by real-world medical applications demonstrate that Continuous-Time Decision Transformer is able to outperform competitors and has clinical utility in terms of improving patients’ health and prolonging their survival by learning high-performance policies from logged data generated using policies of different levels of quality.

We propose Bayesian nonparametric Weibull delegate racing (WDR) to fill the gap in interpretable nonlinear survival analysis with competing events, left truncation, and time-varying covariates. We set a two-phase race among a potentially infinite number of sub-events to model nonlinear covariate effects, which does not rely on transformations or complex functions of the covariates. Using gamma processes, the nonlinear capacity of WDR is parsimonious and data-adaptive. In prediction accuracy, WDR dominates cause-specific Cox and Fine-Gray models and is comparable to random survival forests in the presence of time-invariant covariates. More importantly, WDR can cope with different types of censoring, missing outcomes, left truncation, and time-varying covariates, on which other nonlinear models, such as the random survival forests, Gaussian processes, and deep learning approaches, are largely silent. We develop an efficient MCMC algorithm based on Gibbs sampling. We analyze biomedical data, interpret disease progression affected by covariates, and show the potential of WDR in discovering and diagnosing new diseases.

Motivation: The Cancer Genome Atlas (TCGA) program has produced huge amounts of cancer genomics data providing unprecedented opportunities for research. In 2014, we developed TCGA-Assembler, a software pipeline for retrieval and processing of public TCGA data. In 2016, TCGA data were transferred from the TCGA data portal to the Genomic Data Commons (GDCs), which is supported by a different set of data storage and retrieval mechanisms. In addition, new proteomics data of TCGA samples have been generated by the Clinical Proteomic Tumor Analysis Consortium (CPTAC) program, which were not available for downloading through TCGA-Assembler. It is desirable to acquire and integrate data from both GDC and CPTAC. Results: We develop TCGA-assembler 2 (TA2) to automatically download and integrate data from GDC and CPTAC. We make substantial improvement on the functionality of TA2 to enhance user experience and software performance. TA2 together with its previous version have helped more than 2000 researchers from 64 countries to access and utilize TCGA and CPTAC data in their research. Availability of TA2 will continue to allow existing and new users to conduct reproducible research based on TCGA and CPTAC data. Availability and implementation: http://www.compgenome.org/TCGA-Assembler/ or https://github.com/compgenome365/TCGA-Assembler-2. Contact: [email protected] or [email protected]. Supplementary information: Supplementary data are available at Bioinformatics online.

SUMMARY: Chromosomal copy number variation (CNV) refers to a polymorphism that a DNA segment presents deletion or duplication in the population. The computational algorithms developed to identify this type of variation are usually of high computational complexity. Here we present a user-friendly R package, modSaRa, designed to perform copy number variants identification. The package is developed based on a change-point based method with optimal computational complexity and desirable accuracy. The current version of modSaRa package is a comprehensive tool with integration of preprocessing steps and main CNV calling steps. AVAILABILITY AND IMPLEMENTATION: modSaRa is an R package written in R, C ++ and Rcpp and is now freely available for download at http://c2s2.yale.edu/software/modSaRa . CONTACT: [email protected]. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

A single-nucleotide polymorphism (SNP) is a sole base change in the DNA sequence and is the most common polymorphism. Detection and annotation of SNPs are among the central topics in biomedical research as SNPs are believed to play important roles on the manifestation of phenotypic events, such as disease susceptibility. To take full advantage of the next-generation sequencing (NGS) technology, we propose a Bayesian approach, BM-SNP, to identify SNPs based on the posterior inference using NGS data. In particular, BM-SNP computes the posterior probability of nucleotide variation at each covered genomic position using the contents and frequency of the mapped short reads. The position with a high posterior probability of nucleotide variation is flagged as a potential SNP. We apply BM-SNP to two cell-line NGS data, and the results show a high ratio of overlap ( >95 percent) with the dbSNP database. Compared with MAQ, BM-SNP identifies more SNPs that are in dbSNP, with higher quality. The SNPs that are called only by BM-SNP but not in dbSNP may serve as new discoveries. The proposed BM-SNP method integrates information from multiple aspects of NGS data, and therefore achieves high detection power. BM-SNP is fast, capable of processing whole genome data at 20-fold average coverage in a short amount of time.

MOTIVATION: The phenotypes of knockout mice provide crucial information for understanding the biological functions of mammalian genes. Among various knockout phenotypes, lethality is of great interest because those involved genes play essential roles. With the availability of large-scale genomic data, we aimed to assess how well the integration of various genomic features can predict the lethal phenotype of single-gene knockout mice. RESULTS: We first assembled a comprehensive list of 491 candidate genomic features derived from diverse data sources. Using mouse genes with a known phenotype as the training set, we integrated the informative genomic features to predict the knockout lethality through three machine learning methods. Based on cross-validation, our models could achieve a good performance (accuracy = 73% and recall = 63%). Our results serve as a valuable practical resource in the mouse genetics research community, and also accelerate the translation of the knowledge of mouse genes into better strategies for studying human disease.