--- title: "AnVILWorkflow: Run batch analysis workflows including non-R tools leveraing Cloud resources" author: "Sehyun Oh" date: "`r format(Sys.time(), '%B %d, %Y')`" vignette: > %\VignetteEngine{knitr::rmarkdown} %\VignetteIndexEntry{Quickstart - RNAseq analysis using salmon} %\VignetteEncoding{UTF-8} output: BiocStyle::html_document: number_sections: yes toc: yes toc_depth: 4 --- ```{r setup, include = FALSE} knitr::opts_chunk$set( comment = "#>", collapse = TRUE, fig.align = 'center', eval = AnVIL::gcloud_exists() ) ``` # Overview The [AnVIL project][] is an analysis, visualization, and informatics cloud-based space for data access, sharing and computing across large genomic-related data sets. For R users with the limited computing resources, we introduce the AnVILWorkflow package. This package allows users to run workflows implemented in [Terra][] without installing software, writing any workflow, or managing cloud resources. Terra is a cloud-based genomics platform and its computing resources rely on Google Cloud Platform (GCP). Use of this package requires AnVIL and Google cloud computing billing accounts. Consult [AnVIL training guides][] for details on establishing these accounts. [AnVIL project]: https://anvilproject.org/ [Terra]: https://anvil.terra.bio/ [AnVIL training guides]: https://anvilproject.org/learn ## Install and load package ```{r eval = FALSE} if (!require("BiocManager")) install.packages("BiocManager") BiocManager::install("AnVILWorkflow") ``` ```{r results="hide", message=FALSE, warning=FALSE} library(AnVIL) library(AnVILGCP) library(AnVILWorkflow) ``` ## Google Cloud SDK If you use AnVILWorkflow within Terra's RStudio, you don't need extra authentication and gcloud SDK. If you use this package locally, it requires gcloud SDK and the billing account used in Terra. You can [install][] the gcloud sdk. [isntall]: https://cloud.google.com/sdk/docs/install Check whether your system has the installation with `AnVIL::gcloud_exists()`. It should return `TRUE` to use AnVILWorkflow package. ```{r} gcloud_exists() ``` If it returns `FALSE`, install the gcloud SDK following this script: ```{bash eval=FALSE, echo=FALSE} ## shell $ curl -sSL https://sdk.cloud.google.com | bash ``` ```{r eval=FALSE} devtools::install_github("rstudio/cloudml") cloudml::gcloud_install() ``` ```{bash eval=FALSE} ## shell $ gcloud auth login ``` ```{bash eval=FALSE, echo=FALSE} ## You can change the project using this script ## shell $ gcloud config set project PROJECT_ID ``` ## Create Terra account You need [Terra account setup](https://support.terra.bio/hc/en-us/articles/360034677651-Account-setup-and-exploring-Terra). Once you have your own Terra account, you need two pieces of information to use AnVILWorkflow package: 1) The email address linked to your Terra account\ 2) Your billing project name You can setup your working environment using `setCloudEnv()` function like below. **Provide the input values with YOUR account information!** ```{r eval=FALSE} accountEmail <- "YOUR_EMAIL@gmail.com" billingProjectName <- "YOUR_BILLING_ACCOUNT" setCloudEnv(accountEmail = accountEmail, billingProjectName = billingProjectName) ``` ```{r echo=FALSE} ## In case the environment is set already. accountEmail <- AnVILGCP::gcloud_account() billingProjectName <- AnVILGCP::gcloud_project() setCloudEnv(accountEmail = accountEmail, billingProjectName = billingProjectName, message = FALSE) ``` The remainder of this vignette assumes that an Terra account has been established and successfully linked to a Google cloud computing billing account. ## Major steps Here is the table of major functions for three workflow steps - prepare, run, and check result. | Steps | Functions | Description | |---------|---------------------|-------------------------------------------| | Prepare | `cloneWorkspace` | Copy the template workspace | | | `updateInput` | Take user's inputs | | Run | `runWorkflow` | Launch the workflow in Terra | | | `stopWorkflow` | Abort the submission | | | `monitorWorkflow` | Monitor the status of your workflow run | | Result | `getOutput` | List or download your workflow outputs | ## Example in this vignette: bulk RNAseq analysis You can find all the available workspaces you have access to using `AnVIL::avworkspaces()` function. Workspaces manually curated by this package are separately checked using `availableAnalysis()` function. The values under `analysis` column can be used for the analysis argument, simplifying the cloning process. For this vignette, we use `"salmon"`. ``` > availableAnalysis() analysis workspaceNamespace workspaceName configuration_namespace configuration_name 1 bioBakery waldronlab-terra-rstudio mtx_workflow_biobakery_version3_template mtx_workflow_biobakery_version3 mtx_workflow_biobakery_version3 2 salmon bioconductor-rpci-anvil Bioconductor-Workflow-DESeq2 bioconductor-rpci-anvil AnVILBulkRNASeq 3 pathml waldronlab-terra-rstudio pathml_stain_normalization_template PathML Preprocessing description 1 Microbiome analysis using bioBakery 2 Trascript quantification from RNAseq using Salmon | Differential gene expression analysis using DESeq2 3 Stain normalization step of PathML pipeline ``` ```{r} analysis <- "salmon" ``` # Browse AnVIL resources ```{r} AnVILBrowse("malaria") AnVILBrowse("resistance") AnVILBrowse("resistance", searchFrom = "workflow") ``` # Setup ## Clone workspace ### Curated by this package We will refer the existing workspaces, that you have access to and want to use for your analysis, as 'template' workspaces. The first step of using this package is cloning the template workspace using `cloneWorkspace` function. Note that you need to provide a **unique** name for the cloned workspace through `workspaceName` argument. Once you successfully clone the workspace, the function will return the name of the cloned workspace. For example, the successfully execution of the below script will return `{YOUR_BILLING_ACCOUNT}/salmon_test`. ```{r echo=FALSE, eval=FALSE, error=TRUE} # If you attempt to clone the template workspace using the existing # workspaceName, you will get the error message. cloneWorkspace(workspaceName = "salmon_test", analysis = analysis) ``` ```{r} salmonWorkspaceName <- basename(tempfile("salmon_")) # unique workspace name salmonWorkspaceName cloneWorkspace(workspaceName = salmonWorkspaceName, analysis = analysis) ``` ```{r cleanup, echo=FALSE, message=FALSE, error=TRUE, warning=FALSE} ## Delete test workspaces resp <- AnVIL::Terra()$deleteWorkspace(workspaceNamespace = billingProjectName, workspaceName = salmonWorkspaceName) rm(resp) ``` ### Any workspace you have access to If you want to clone any other workspace that you have access to but is not curated by this pacakge, you can directly enter the name of the target workspace as a `templateName`. For example, to clone the [Tumor_Only_CNV][] workspace: [Tumor_Only_CNV]: https://anvil.terra.bio/#workspaces/waldronlab-terra/Tumor_Only_CNV ```{r eval=FALSE} cnvWorkspaceName <- basename(tempfile("cnv_")) # unique workspace name cnvWorkspaceName cloneWorkspace(workspaceName = cnvWorkspaceName, templateName = "Tumor_Only_CNV") ``` ```{r echo=FALSE} ## workspace used in this vignette salmonWorkspaceName <- "salmon_test" ``` ## Prepare input ### Current input You can review the current inputs using `currentInput` function. Below shows all the required and optional inputs for the workflow. ```{r} config <- getWorkflowConfig(workspaceName = salmonWorkspaceName) current_input <- currentInput(salmonWorkspaceName, config = config) current_input ```
### Update input You can modify/update inputs of your workflow using `updateInput` function. To minimize the formatting issues, we recommend to make any change in the current input table returned from the `currentInput` function. Under the default (`dry=TRUE`), the updated input table will be returned without actually updating Terra/AnVIL. Set `dry=FALSE`, to make a change in Terra/AnVIL. ```{r} new_input <- current_input new_input[4,4] <- "athal_index" new_input updateInput(salmonWorkspaceName, inputs = new_input, config = config) ``` # Run workflow You can launch the workflow using `runWorkflow()` function. You need to specify the `inputName` of your workflow. If you don't provide it, this function will return the list of input names you can use for your workflow. Example error outputs: ```{r eval=FALSE} runWorkflow(slamonWorkspaceName, config = config) # You should provide the inputName from the followings: # [1] "AnVILBulkRNASeq_set" #> Error in runWorkflow(salmonWorkspaceName): ``` ```{r} runWorkflow(salmonWorkspaceName, inputName = "AnVILBulkRNASeq_set", config = config) ``` ## Monitor progress The last three columns (`status`, `succeeded`, and `failed`) show the submission and the result status. ```{r} submissions <- monitorWorkflow(workspaceName = salmonWorkspaceName) submissions ``` ## Abort submission You can abort the most recently submitted job using the `stopWorkflow` function. You can abort any workflow that is not the most recently submitted by providing a specific `submissionId`. ```{r} stopWorkflow(salmonWorkspaceName) ``` # Result The workspace `Bioconductor-Workflow-DESeq2` is the template workspace you cloned at the beginning using the `analysis = "salmon"` argument in `cloneWorkspace()` function. This template workspace has already a history of the previous submissions, so we will check the output examples in this workspace. ```{r} submissions <- monitorWorkflow(workspaceName = "Bioconductor-Workflow-DESeq2") submissions ``` You can check all the output files from the most recently succeeded submission using `getOutput` function. If you specify the `submissionId` argument, you can get the output files of that specific submission. ```{r no-run-examples, eval=FALSE} ## Output from the successfully-done submission successful_submissions <- submissions$submissionId[submissions$succeeded == 1] out <- getOutput(workspaceName = "Bioconductor-Workflow-DESeq2", submissionId = successful_submissions[1]) ``` ```{r echo=FALSE} ## Save the previous submission results from the above chunk ## write.table(out, "vignettes/salmon_test_outputs.csv") out <- read.table("salmon_test_outputs.csv", header = TRUE) %>% tibble::tibble() ``` ```{r} head(out) ``` # Session Info ```{r eval=TRUE} sessionInfo() ```