Episode: Working with genomic file formats

[kevinrue]

The list below is a suggestion of individual items in this episode.

Feedback is welcome!

Ticked boxes indicate items that already have a dedicated section in this episode.

Lesson points

• rtracklayer
• Import data from common file formats
• Export data from common file formats
• Class of S4 object compatible with each file format

Challenges

• Import/export data from/to various common file formats

Notes

• This episode partially overlaps with other episodes that demonstrate the rtracklayer package in specific contexts (e.g., Episode: Working with genomic ranges #27 )

[kevinrue]

Having worked on the "genomic ranges" episode in the meantime, I wonder whether it is worth to have a separate episode that would attempt to cover multiple file formats and rtracklayer separately. The lesson about genomic ranges demonstrates how to import a GTF file as a GRanges object.

This begs the questions:

• Would the demonstration of rtracklayer::import() need to move out of the "genomic ranges" episode, and into an "rtracklayer" episode, which would then have to be inserted somewhere before the "genomic ranges" episode (and any other episode that refers to "rtracklayer"?

  • negative: rtracklayer episode might become huge as more episodes are added; rtracklayer episode might be very abstract if it is only about importing from and exporting to files without the context of a workflow in-between; rtracklayer would be one of the first episodes in the lesson, introducing import/export before the actual data structures that it relies on
  • positive: perhaps demonstrating import/export without the distraction of an analysis workflows might be useful (?)

• Would we duplicate the demonstration of rtracklayer::import() in an "rtracklayer" episode as well as any other episode that uses rtracklayer ?

  • negative: repetition and possibly maintenance
  • positive: reduce dependencies between episodes and make them standalone units

• Should we simply drop the idea of a separate "rtracklayer" episode, and mention the relevant import/export functionality in each episode that needs it?

  • negative: dilutes the visibility of the rtracklayer package over multiple episodes
  • positive: avoids repetition and inter-dependencies between episode that would complicate the maintenance of the topic