Recommended Quality Control (QC) pipeline for alevin-fry input

[lilyzhou222]

Hi alevin-fry developers,

First, thank you for developing such a powerful and efficient tool for single-cell analysis.

I am currently setting up a pipeline to process single-cell RNA-seq data to quantify spliced and unspliced abundances, and I plan to use alevin-fry for this purpose. I have a question regarding the recommended quality control (QC) steps and their placement in the workflow. I've reviewed the documentation but couldn't find a specific section detailing best practices for pre-processing the raw FASTQ files before they are used to generate the RAD file.

Could you please clarify the following points?

Necessity of QC: Is it generally recommended to perform standard QC on the raw FASTQ files (e.g., using tools like fastp or Trimmomatic) before processing them with alevin to generate the RAD file?

Timing of QC: If QC is recommended, should it be performed before generating the RAD file? My assumption is that pre-processing the FASTQs is the correct approach, but I would like to confirm this.

Specific QC Steps: If pre-processing is needed, which QC steps are most critical for ensuring accurate quantification with alevin-fry? For example:

Adapter trimming?

Read quality filtering (e.g., by mean quality score)?

Length filtering?

Any guidance or clarification on the best practices for input data quality would be greatly appreciated. This information could also be a valuable addition to the documentation for new users.

Thank you for your time and help!