Creating BAM from the fastq files

[hiraksarkar]

Hi Developers,

Thanks for putting up a great pipeline. I am trying to create alignment bam files (locally at hms outside broad) from the newly published paper. The raw fastq files are mentioned in the paper.
I have prepared the reference and currently looking for ways to run /src/slideseq/scripts/demultiplex.sh, unfortunately, I am not able to follow some of the arguments in this script.

Is there a place where I can receive more information about the run, except the raw fastq files?

Best,
Hirak

[hiraksarkar]

Hi Developers,

Any hint on the above queries would be really helpful.

[jamestwebber]

Apologies for not seeing this earlier, I haven't been getting notifications for issues here.

Hopefully this is resolved with our discussion today. For others searching the repo: the slideseq pipeline does not use fastq files directly, but instead uses unaligned BAM files (commonly used at the Broad) for most operations.

[ruqianl]

Hi @jamestwebber,

Given the fastq files shared via SRA from the paper mentioned in the original post, so do we turn that into uBAM first to apply slide-seq-tools.
What function to call? I cloned the repo and created the conda env, not sure how to supply the inputs with:

Usage: submit_slideseq [OPTIONS] RUN [RUN ...]

  Submit each RUN to the Slide-seq alignment pipeline.

  See README.md for instructions and requirements:
  github.com/MacoskoLab/slideseq-tools

Options:
  --demux / --no-demux            Whether to run demultiplexing
  --align / --no-align            Whether to run alignment
  --processing / --no-processing  Whether to run processing
  --dryrun                        Show the plan but don't execute
  --debug                         Turn on debug logging
  --log-file PATH                 File to write logs
  --help                          Show this message and exit.

Is this tool currently intended to work for external users? (external to Broad).

Thanks.
Ruqian

[jamestwebber]

This pipeline starts from BCL files (the raw output of an Illumina sequencer) and creates unaligned BAM files. There's no mechanism to input fastq because we don't use them.

Is this tool currently intended to work for external users? (external to Broad).

As mentioned in the README, the pipeline is pretty specialized for our current workflow. There are many tradeoffs between building a general tool and running things efficiently and (as far as I'm aware) the vast majority of Slide-seq processing is done in the lab. I tried to make the pipeline reasonably easy to read, but running in another environment will take some work.

For instance, we use a Google Sheet to keep track of experiments, and the tool is built to retrieve information from that sheet (the location of the sheet is at least configurable). So the command line tool is built to fetch the sheet and look for runs based on the RUN_NAME arguments.

I will also mention: there is a team here working on a Terra workflow for Slide-seq analysis, which might be easier to use outside of the Broad.

[ruqianl]

thanks for the reply. The uploaded files from a study that is available to public are generally fastqs not BCL's. But I understand that the pipeline is specialized to internal users with raw sequencing files to process.

[jamestwebber]

The uploaded files from a study that is available to public are generally fastqs not BCL's.

Yes, I'm well aware!

Due to the infrastructure involved it would be difficult to reproduce the published analysis exactly, but the basic pieces of this pipeline are fairly simple: R1 contains the barcode and UMI, and R2 contains the cDNA read. So we align R2 to the reference, match them to beads with the barcode in R1, and count the UMIs. There is a small amount of custom logic in the bead-matching step (very similar beads in close proximity are combined), but that step is not critical for getting good results.

The Terra workflow I mentioned above will take fastqs as input, so once that is available it should be easier for others to run analysis--even if you do not want to use Terra, the workflow will be useful.