Search

Genome-wide association studies (GWAS) are among the workhorses of statistical genetics, having detected thousands of variants associated with complex traits and diseases. A typical GWAS examines the association between genotypes and the phenotype of interest while adjusting for a set of covariates. While covariates potentially have non-linear effects on the phenotype in many real world settings, due to the challenge of specifying the model, GWAS seldom include non-linear terms. Here we introduce DeepNull, a method that models non-linear covariate effects on phenotypes using a deep neural network (DNN) and then includes the model prediction as a single extra term in the GWAS association. First, using simulated data, we show that DeepNull increases statistical power by up to 20% while maintaining tight control of the type I error in the presence of interactions or non-linear covariate effects. Second, DeepNull maintains similar results to a standard GWAS when covariates have only linear effects on the phenotype. Third, DeepNull detects larger numbers of significant hits and loci (7% additional loci averaged over 10 traits) than standard GWAS in ten phenotypes from the UK Biobank (n=370K). Many of the hits found only by DeepNull are biologically plausible or have previously been reported in the GWAS catalog. Finally, DeepNull improves phenotype prediction by 23% averaged over the same ten phenotypes, the highest improvement was observed in the case of Glaucoma referral probability where DeepNull improves the phenotype prediction by 83%.

Genome-wide association studies (GWASs) examine the association between genotype and phenotype while adjusting for a set of covariates. Although the covariates may have non-linear or interactive effects, due to the challenge of specifying the model, GWAS often neglect such terms. Here we introduce DeepNull, a method that identifies and adjusts for non-linear and interactive covariate effects using a deep neural network. In analyses of simulated and real data, we demonstrate that DeepNull maintains tight control of the type I error while increasing statistical power by up to 20% in the presence of non-linear and interactive effects. Moreover, in the absence of such effects, DeepNull incurs no loss of power. When applied to 10 phenotypes from the UK Biobank (n = 370K), DeepNull discovered more hits (+6%) and loci (+7%), on average, than conventional association analyses, many of which are biologically plausible or have previously been reported. Finally, DeepNull improves upon linear modeling for phenotypic prediction (+23% on average)., GWAS often assume a linear phenotype-covariate relationship which may not hold in practice. Here the authors present DeepNull, in which they apply deep learning to identify and adjust for complex non-linear relationships, improving phenotypic prediction and GWAS power.

Despite rapid advances in sequencing technologies, accurately calling genetic variants present in an individual genome from billions of short, errorful sequence reads remains challenging. Here we show that a deep convolutional neural network can call genetic variation in aligned next-generation sequencing read data by learning statistical relationships between images of read pileups around putative variant and true genotype calls. The approach, called DeepVariant, outperforms existing state-of-the-art tools. The learned model generalizes across genome builds and mammalian species, allowing nonhuman sequencing projects to benefit from the wealth of human ground-truth data. We further show that DeepVariant can learn to call variants in a variety of sequencing technologies and experimental designs, including deep whole genomes from 10X Genomics and Ion Ampliseq exomes, highlighting the benefits of using more automated and generalizable techniques for variant calling.

Gradient boosted decision trees are a popular machine learning technique, in part because of their ability to give good accuracy with small models. We describe two extensions to the standard tree boosting algorithm designed to increase this advantage. The first improvement extends the boosting formalism from scalar-valued trees to vector-valued trees. This allows individual trees to be used as multiclass classifiers, rather than requiring one tree per class, and drastically reduces the model size required for multiclass problems. We also show that some other popular vector-valued gradient boosted trees modifications fit into this formulation and can be easily obtained in our implementation. The second extension, layer-by-layer boosting, takes smaller steps in function space, which is empirically shown to lead to a faster convergence and to a more compact ensemble. We have added both improvements to the open-source TensorFlow Boosted trees (TFBT) package, and we demonstrate their efficacy on a variety of multiclass datasets. We expect these extensions will be of particular interest to boosted tree applications that require small models, such as embedded devices, applications requiring fast inference, or applications desiring more interpretable models., Comment: Accepted for publication in IEEE Big Data 2017 http://cci.drexel.edu/bigdata/bigdata2017/AcceptedPapers.html

TF Boosted Trees (TFBT) is a new open-sourced framework for the distributed training of gradient boosted trees. It is based on TensorFlow, and its distinguishing features include a novel architecture, automatic loss differentiation, layer-by-layer boosting that results in smaller ensembles and faster prediction, principled multi-class handling, and a number of regularization techniques to prevent overfitting.

Next-generation sequencing (NGS) is a rapidly evolving set of technologies that can be used to determine the sequence of an individual’s genome1 by calling genetic variants present in an individual using billions of short, errorful sequence reads2. Despite more than a decade of effort and thousands of dedicated researchers, the hand-crafted and parameterized statistical models used for variant calling still produce thousands of errors and missed variants in each genome3,4. Here we show that a deep convolutional neural network5 can call genetic variation in aligned next-generation sequencing read data by learning statistical relationships (likelihoods) between images of read pileups around putative variant sites and ground-truth genotype calls. This approach, called DeepVariant, outperforms existing tools, even winning the “highest performance” award for SNPs in a FDA-administered variant calling challenge. The learned model generalizes across genome builds and even to other mammalian species, allowing non-human sequencing projects to benefit from the wealth of human ground truth data. We further show that, unlike existing tools which perform well on only a specific technology, DeepVariant can learn to call variants in a variety of sequencing technologies and experimental designs, from deep whole genomes from 10X Genomics to Ion Ampliseq exomes. DeepVariant represents a significant step from expert-driven statistical modeling towards more automatic deep learning approaches for developing software to interpret biological instrumentation data.

This paper describes the progress made in the transcription of broadcast news (BN) and conversational telephone speech (CTS) within the combined BBN/LIMSI system from May 2002 to September 2004. During that period, BBN and LIMSI collaborated in an effort to produce significant reductions in the word error rate (WER), as directed by the aggressive goals of the Effective, Affordable, Reusable, Speech-to-text [Defense Advanced Research Projects Agency (DARPA) EARS] program. The paper focuses on general modeling techniques that led to recognition accuracy improvements, as well as engineering approaches that enabled efficient use of large amounts of training data and fast decoding architectures. Special attention is given on efforts to integrate components of the BBN and LIMSI systems, discussing the tradeoff between speed and accuracy for various system combination strategies. Results on the EARS progress test sets show that the combined BBN/LIMSI system achieved relative reductions of 47% and 51% on the BN and CTS domains, respectively

In this paper we describe the English Conversational Telephone Speech (CTS) recognition system jointly developed by BBN and LIMSI under the DARPA EARS program for the 2004 evaluation conducted by NIST. The 2004 BBN/LIMSI system achieved a word error rate (WER) of 13.5% at 18.3xRT (realtime as measured on Pentium 4 Xeon 3.4 GHz Processor) on the EARS progress test set. This translates into a 22.8% relative improvement in WER over the 2003 BBN/LIMSI EARS evaluation system, which was run without any time constraints. In addition to reporting on the system architecture and the evaluation results, we also highlight the significant improvements made at both sites.

We report on the results of the first evaluations for the BBN/LIMSI system under the new DARPA EARS program. The evaluations were carried out for conversational telephone speech (CTS) and broadcast news (BN) for three languages: English, Mandarin, and Arabic. In addition to providing system descriptions and evaluation results, the paper highlights methods that worked well across the two domains and those few that worked well on one domain but not the other. For the BN evaluations, which had to be run under 10 times real-time, we demonstrated that a joint BBN/LIMSI system with a time constraint achieved better results than either system alone.

This paper describes the BYBLOS system that BBN used to participate in the 2001 NIST Hub-5 evaluation benchmark. We outline the procedure used for training and decoding, and present the algorithmic improvements made to the system, along with experimental results. These improvements include a Gaussian splitting initialization procedure, the use of Linear Discriminant Analysis, and processing of additional acoustic training data. We also discuss our system combination and confidence-based thresholding methods. Experiments on an internal validation test set show that all these system improvements provide a 8.1 % relative reduction in word error rate compared to our 2000 LVCSR system.