Working with Bascet-FASTQ

Advanced users may wish to process the reads using custom software. As such software typically ingests FASTQ-files, we provide Bascet-FASTQ. These files are regular (pairwise FASTQ-files with a naming scheme).

There are a few things you can do with these: * Special trimming (but Bascet already performs rudimentary trimming) * Use alternative aligners (the output BAM files retains read names and can thus be ingested by Bascet again) * Extract special sequences for separate analysis (perturb-seq sgRNAs; CITE-seq; etc) * Deposit the FASTQ into archives (but read separate section!)

We assume that you have set up for Instance and Runner for the following commands to work.

Raw FASTQ depositing

Bascet-FASTQ is possibly suitable for depositing raw data. The data will take less space as trimming and barcode identification has already been performed. The sorting also improves compression ratio. The presorting further reduces work for reproducing the analysis.

However, there is a major caveat: FASTQ-files submitted to SRA will have the names of reads removed in their “light” format (i.e., which cell the read belongs to). It will still be possible to get the fully raw files using the cold storage retrieval option, but users may have to pay for this, and download takes longer times.

We are working toward a solution, but users are currently better off submitting raw data to Zenodo or other general depository. In such a case, you might as well submit the native TIRP files, but we currently do not make any recommendation.

Conversion to Bascet-FASTQ

The following command converts any Bascet file containing reads to Bascet-FASTQ:

### Get reads in FASTQ format
BascetMapTransform(
  bascetRoot,
  inputName="filtered",
  outputName="asfq",
  outFormat="R1.fq.gz"
)

Trimming with FASTP

(FASTP) is a fast trimmer. We however don’t recommend it for de novo assembly as it seems to leave too many adapters - this is why we already perform our own trimming during GetRaw. In either case, FASTP can be run directly through Zorn. If you wish to use other software, you can apply it in any other way to the FASTQ files.

### Get reads in fastq format
BascetRunFASTP(
  bascetRoot,
  numLocalThreads=10,
  inputName="asfq",
  outputName="fastp"
)

Conversion from Bascet-FASTQ to TIRP

We recommend keeping reads in TIRP files as this is what commands are optimized for; and TIRP offers direct extraction of reads from a cell, via indexing, unlike FASTQ. The following command performs the conversion

#Convert FASTQ to TIRP
BascetMapTransform(
  bascetRoot,
  "fastp",
  "new_filtered",
  outFormat="tirp.gz"
)

You now have a TIRP file again, equivalent to the first shardified output. !!! Don’t forget to specify inputName=“new_filtered” to later commands, as the default is to use “filtered”