Matching cluster label to origin sample id

[elemesemele]

Dear @wheaton5, thank you for developing souporcell!
I want to match the cluster labels in the Souporcell output (clusters.tsv) to the original sample IDs.

I ran the pipeline using --known_genotypes and --known_genotypes_sample_names (with --skip_remap True).

The column in the --known_genotypes vcf file has three sample IDs (SampleA, SampleB, samlpeC) and I entered --known_genotypes_sample_names SampleA SampleB SampleC in the command.

And the singlecell GEM was actually multiplexed into these three samples (SampleA, SampleB, SampleC).

Is the order of the cluster labels in 'clusters.tsv' the same as the order of the sample IDs I entered (0=SampleA, 1=SampleB, 2=SampleC)?

thank you!!

[wheaton5]

Yes that should be correct

[elemesemele]

Yes that should be correct

Thank you for your kind response.
It's really nice that the order of the cluster is the same as the order of the original samples entered as input.

But is there any possibility that this matching is wrong?

For testing purposes, I ran the above command without skipping remapping (--known_genotypes and --known_genotypes_sample_names (with --skip_remap False).

As a result, the assigned cluster number has changed.
Why is this?

These are my results.
I think the order of cluster0 and cluster1 has been swapped.

--skip_remap True

--skip_remap False

Would you recommend the --skip_remap True option to assign cluster order and original samples accordingly?

Sincerely.

[plijnzaad]

If all is well results should be nearly identical. The --skip_remap just avoids the needless and time-consuming remapping (reassembling, really) of all the reads.