kharchenkolab/dropEst

Use two fastq files for 10x in dropTag

Opened this issue · 5 comments

Hi, I am trying to apply your pipeline on my 10X Genomics data. When I run command for dropTaq. It didn't work for me. I don't have separate 3 files in my data as you explained in your pipeline. I am also attaching the files that I received from 10X.
screen shot 2018-09-18 at 16 47 18

I also tried to get help from dropTag but I received this error.
[sikh@login-0 dropEst]$ droptag -h
-bash: droptag: command not found

How I can fix this issue. Thanks in advance.

I have the same problem, I would like to use droptag on data from 10X data of single cell TCR enriched data (VDJ). Is that possible?

Hi,
10x uses complicated way of library barcoding, so I'd recommend to use CellRanger instead of droptag. And then, you can still use dropest on the bam file, produced by CellRanger. Can I ask you why do you prefer to use droptag?

@amarquard , never tried that. What's special in this data in terms of processing?

It also doen't have the 3 fastq files you describe in the pipeline. It has read 1 (which includes cell barcode and UMI but also has transcript), read 2 (only transcript) and library tag. I would like to use droptag instead of cellranger because I am using a customised 10x setup, and thought perhaps droptag would be more suited to fit in my own analysis pipeline than cellranger. But I could be wrong. Do you have more information somewhere on how exactly droptag works, its options and arguments etc? thanks in advance.

Now if you have 10x without library tag, you can run it by using Split-seq config with adjusted parameters. To do so, add the following to config/TagsSearch:

    <protocol>split_seq</protocol>
    <MultipleBarcodeSearch>
        <barcode_starts>0</barcode_starts>
        <barcode_lengths>16</barcode_lengths>
        <umi_start>16</umi_start>
        <umi_length>26</umi_length>
    </MultipleBarcodeSearch>

Don't forget to remove <protocol>10x</protocol> from the config

Fantastic, I look forward to trying it out. Thanks a lot!