search

single-cell-genetics / vireo

Demultiplexing pooled scRNA-seq data with or without genotype reference
https://vireoSNP.readthedocs.io
Apache License 2.0
73 stars 27 forks source link

Different result with same file and cmd #18

Open mengchengyao opened 3 years ago

mengchengyao commented 3 years ago

Hi, Thanks for great tool for donor ananylsis. I got different results with same file and cmd, I do not know why? image

Best

mengchengyao commented 3 years ago

seting --randSeed maybe the same result, however, what randSeed should i set, especially i focus on doublet? I want to detect cell fusion.

Best

huangyh09 commented 3 years ago

Hi, indeed the two runs are a bit different, but probably not too much as it looks like the order of donor1 and donor2 swapped in these two runs. You could set an arbitrary random seed, e.g., 1.

It looks like that you have a high proportion of unassigned cells, which may affect the doublets detection and probably can be improved by changing the settings.

  1. Is it possible to access to the genotype of any of these two donors, e.g., bulk RNA-seq? This usually helps.
  2. It looks you have 3,485 variants. Which candidate SNPs do you use? Maybe you can try a larger candidate list, e.g., the 36 million SNPs with AF>=0.0005.
  3. Are you using cellSNP for genotyping? If so, which version? You should use >= v0.3.1 by turning off the PCR duplicate removal, so more (informative) reads will be kept. Also, you are recommend to use the C/C++ re-implementation cellsnp-lite for higher efficiency.

Yuanhua