Rapid and efficient analysis of 20,000 RNA-seq samples with Toil

Abstract

Toil is portable, open-source workflow software that supports contemporary workflow definition languages and can be used to securely and reproducibly run scientific workflows efficiently at large-scale. To demonstrate Toil, we processed over 20,000 RNA-seq samples to create a consistent meta-analysis of five datasets free of computational batch effects that we make freely available. Nearly all the samples were analysed in under four days using a commercial cloud cluster of 32,000 preemptable cores. Figure 1. (Left)A dependency graph of the RNA-seq pipeline we developed (called CGL). CutAdapt was used to remove extraneous adapters, STAR was used for alignment and read coverage, and RSEM and Kallisto were used to produce quantification data. (Right) A scatter plot showing the Pearson correlation between the results of the TCGA best-practices pipeline and the CGL pipeline. 10,000 randomly selected sample/gene pairs were subset from the entire TCGA cohort and the normalized counts were plot against each other; this process was repeated 5 times with no change in Pearson correlation. The unit for counts is: log2(norm counts+1).Contemporary genomic datasets contain tens of thousands of samples and petabytes of sequencing data 1,2,3 . Genomic processing pipelines can consist of dozens of individual steps, each with their own set of parameters 4,5 . As a result of this size and complexity, computational resource limitations and reproducibility are becoming a major concern within genomics. In response to these interrelated issues, we have created Toil. Reproducible WorkflowsTo support the sharing of scientific workflows, Toil is the first software to execute Common Workflow Language (CWL, Supplementary Note 7) and provide draft support for Workflow Description Language (WDL), both 1 . CC-BY-NC 4.0 International license It is made available under a was not peer-reviewed) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.The copyright holder for this preprint (which . http://dx.doi.org/10.1101/062497 doi: bioRxiv preprint first posted online Jul. 7, 2016; burgeoning standards for scientific workflows 6,7 . A workflow is composed of a set of tasks, or jobs, that are orchestrated by specification of a set of dependencies that map the inputs and outputs between jobs. In addition to CWL and draft WDL support, Toil provides a Python API that allows workflows to be declared statically, or generated dynamically, so that jobs can define further jobs as needed (Supplementary Note 1). The jobs defined in either CWL or Python can consist of Docker containers, which permit sharing of a program without requiring individual tool installation or configuration within a specific environment. Open-source workflows that invoke containers can therefore be run precisely and reproducibly, regardless of environment. We provide a repository of workflows as examples 8 . Toil also integrates with Apache Spark 9 (Supplementary Note 6, Supplementary Fig. 4), and can be used to rapidly create cont...

Identification of GPC2 as an Oncoprotein and Candidate Immunotherapeutic Target in High-Risk Neuroblastoma

“…The TARGET neuroblastoma, other pediatric tumor RNA sequencing data and GTEx RNA sequencing data was downloaded from S3 buckets (Amazon; s3://cgl-rnaseq-recompute-fixed/target/ and s3://cgl-rnaseq-recompute-fixed/gtex/) on 8/5/2016 from prior processed data as described in detail below from the UCSC Computational Genomics Laboratory (Vivian et al, 2016). …”

Developing a framework for digital objects in the Big Data to Knowledge (BD2K) commons: Report from the Commons Framework Pilots workshop

“…Docker containers that are coded in CWL and WDL facilitate scalable, efficient, and reproducible deployment of tools across platforms including cloud environments. In addition, the BD2K Genomics Center has developed Toil [36] (Table 2, Standards category), and the BDDS center has developed Globus Genomics [37] (Table 2, Computing Platform category). Similar to Cromwell, Nextflow, and Arvados (Table 2, Computing Platform category), the aim of Toil and Globus Genomics is to make it easier for users to run large-scale analyses.…”

Rabix: An Open-Source Workflow Executor Supporting Recomputability and Interoperability of Workflow Descriptions

“…It is important to note that the “parent-child” terminology is also applied to relations between individual workflow nodes by the Toil project, an executor which can also interpret Common Workflow Language. 10 However, Rabix uses these terms to refer to computational "jobs" and "subjobs", e.g. a “nested” workflow node is a child of a workflow and can be decomposed into an array of “subjobs”.…”

“…As compared to other CWL execution models 2,10 , computational events are triggered by “port” events instead of “job” events. In other words, when a port is evaluated, this triggers the executor to scan or update these tables in the following order: Variables, Jobs, Links.…”

“…However, certain features described here are also used by existing workflow systems, 6,7,10–12 most notably the support for multiple infrastructures. Additionally, there are certain features not yet implemented in Rabix but which are seen in other systems, such as conditional steps in a workflow, as seen in Toil.…”