Skip to content

Latest commit

 

History

History
45 lines (35 loc) · 2.12 KB

sort-filepairs-by-samplename.md

File metadata and controls

45 lines (35 loc) · 2.12 KB

Problem

You have many tuples with a sample name that works as an identifier and you want to sort the channel by the sample name.

Solution

Use the toSortedList and the flatMap operators to convert the channel to a sorted list, and then convert back to the original structure you get from the fromFilePairs channel factory.

It's worth mentioning that the toSortedList operator is not scalable as it introduces a blocking point in the pipeline execution since to sort the elements of a channel, it needs to collect all of them first.

The fromFilePairs channel factory in the code below will create a channel with tuple elements on the following format:

[
  [samplec, [/path/to/my/files/samplec_1.fastq, /path/to/my/files/samplec_2.fastq]]
  [sampleb, [/path/to/my/files/sampleb_1.fastq, /path/to/my/files/sampleb_2.fastq]]
  [samplea, [/path/to/my/files/samplea_1.fastq, /path/to/my/files/samplea_2.fastq]]
  [sampled, [/path/to/my/files/sampled_1.fastq, /path/to/my/files/sampled_2.fastq]]
  [samplee, [/path/to/my/files/samplee_1.fastq, /path/to/my/files/samplee_2.fastq]]
]

Code

Channel
  .fromFilePairs('/path/to/my/files/*_{1,2}.fastq')
  // Sort the channel elements based on the first object of each tuple,
  // that is, the sample name, and convert to a channel with a single
  // element which is a list of tuples
  .toSortedList( { a, b -> a[0] <=> b[0] } ) // <=> is an operator for comparison
  // flatten the single-element channel to a channel with as many elements
  // as there are samples, which is the original structure provided by
  // fromFilePairs
  .flatMap()
  // View the channel elements by printing it to the screen
  .view()

Run it

Run the example using this command:

nextflow run nextflow-io/patterns/sort-filepairs-by-samplename.nf