passing bam files to sc_long_multisample_pipeline

[sparthib]

Hi there,

I have generated BAM files for all my samples manually using minimap2, what would be the format to pass these ontosc_long_multisample_pipeline? I see options listed for passing the FASTQ files to the function in the documentation, is it similar for the bams?

Thanks,
Sowmya

[ChangqingW]

You can copy / symlink the BAM files to the output folder, name them as [corresponding_fastq_file_name]_align2genome.bam, e.g. if you have sample1.fastq and sample2.fastq, then put sample1_align2genome.bam and sample2_align2genome.bam should make FLAMES skip the alignment step and use the provided BAM.
This will also work for realignment (sample1_realign2transcript.bam).

[sparthib]

great thanks!

[sparthib]

is there a similar multi-sample pipeline for bulk samples? Thanks!

Sowmya

[ChangqingW]

For now, you could put all FASTQs into one folder and provide the path to the folder, each FASTQ file would be considered a sample and the [corresponding_fastq_file_name]_align2genome.bam file could skip the alignment step.
The plan is to make this the same as the sc_long_multisample_pipeline in the devel branch where you could simply provide a named vector, where values could be path to folder or file:

sc_long_multisample_pipeline(
fastqs = c(
  "sample1" = file.path(outdir, "fastq", "your_fq_folder_for_1st_sample"),
  "sample2" = file.path(outdir, "fastq", "your_second_fq.fq.gz"),
  "sample3" = file.path(outdir, "fastq", "third.fq.gz")),
...
)

And then the names would be used as sample names

[sparthib]

just for clarification, are you saying sc_long_multisample_pipeline can be used for bulk samples as well?

thanks,
Sowmya

[ChangqingW]

Sorry, I meant that bulk_long_pipeline can handle multiple samples, but you need to put all FASTQs into one folder and provide the path to the folder, each FASTQ file would be considered a sample.

[sparthib]

I tried something like this:

fastq_path = "path_to_input_fastqs"
fastqs <- list.files(fastq_path, full.names = TRUE)

names(fastqs) <- c("sample1", "sample2" ...)

se <- bulk_long_pipeline(
  annotation = gtf, 
  fastq= fastqs,
  outdir = outdir,
  genome_fa = genome_fa, config_file = config_file
)

and it returns an error like this

Error in if (utils::file_test("-d", fastq)) { : 
  the condition has length > 1
Calls: bulk_long_pipeline

I believe this has to do with passing multiple fastq files to the fastq argument but let me know if I'm wrong.

Thanks,
Sowmya

[ChangqingW]

For now you could use

fastq_path = "path_to_input_fastqs"
# fastqs <- list.files(fastq_path, full.names = TRUE)
# names(fastqs) <- c("sample1", "sample2" ...)

se <- bulk_long_pipeline(
  annotation = gtf, 
  fastq= fastq_path,
  outdir = outdir,
  genome_fa = genome_fa, config_file = config_file
)

each FASTQ file would then be considered as a sample.