This feature is only available in organization workspaces.
Datasets in Nextflow Tower are CSV (comma-separated values) and TSV (tab-separated values) formatted files stored in a workspace. They are designed to be used as inputs to pipelines to simplify data management, minimize user data-input errors, and facilitate reproducible workflows.
The most commonly used datasets for Nextflow pipelines are samplesheets, where each row consists of a sample, the location of files for that sample (such as fastq files), and other sample details. For example, nf-core/rnaseq works with input datasets (samplesheets) containing sample names, fastq file locations, and indications of strandedness. The Tower community showcase sample dataset for nf-core/rnaseq looks like this:
???+ example "rnaseq dataset"
The combination of datasets, pipeline secrets, and pipeline actions in Tower allows you to automate workflows to curate your data and maintain and launch pipelines based on specific events. See here for an example of pipeline workflow automation using Tower.
Using datasets reduces errors that occur due to manual data entry when launching pipelines.
Datasets can be generated automatically in response to events (such as S3 storage new file notifications).
Datasets can streamline differential data analysis when using the same pipeline to launch a run for each dataset as it becomes available.
For your pipeline to use your dataset as input during runtime, information about the dataset and file format must be included in the relevant parameters of your pipeline schema. The pipeline schema specifies the accepted dataset file type in the
mimetype attribute (either
Dataset validation and file content requirements
Tower does not validate your dataset file contents. While datasets can contain static file links, you are responsible for maintaining the access to that data.
Datasets can point to files stored in various locations, such as Amazon S3 or GitHub. To stage the file paths defined in the dataset, Nextflow requires access to the infrastructure where the files reside, whether on Cloud or HPC systems. Add the access keys for data sources that require authentication to your pipeline secrets.
All Tower users have access to the datasets feature in organization workspaces.
Creating a new dataset
To create a new dataset, follow these steps:
- Open the Datasets tab in your organization workspace.
- Select New dataset.
- Complete the Name and Description fields using information relevant to your dataset.
- Add the dataset file to your workspace with drag-and-drop or the system file explorer dialog.
- For dataset files that use the first row for column names, customize the dataset view with the First row as header option.
The size of the dataset file cannot exceed 10MB.
Tower can accommodate multiple versions of a dataset. To add a new version for an existing dataset, follow these steps:
- Select Edit next to the dataset you wish to update.
- In the Edit dialog, select Add a new version.
- Upload the newer version of the dataset and select Update.
All subsequent versions of a dataset must be the same format (.csv or .tsv) as the initial version.
Using a dataset
To use a dataset with the saved pipelines in your workspace, follow these steps:
- Open any pipeline that contains a pipeline-schema from the Launchpad.
- Select the input field for the pipeline, removing any default value.
- Pick the dataset to use as input to your pipeline.
The input field drop-down menu will only display datasets that match the file type specified in the
nextflow_schema.json of the chosen pipeline. If the schema specifies
"mimetype": "text/csv", no TSV datasets will be available for use with that pipeline, and vice versa.