Expected Runtime for 550k SNPs, 500k Respondents and 100 PCs

[Tobias Wolfram]

Hi, I'm running FlashPCA2 on an Ubuntu Workstation from the command line for roughly 550k SNPs, 500k Respondents and require the first 100 PCs. So my command is

/data/flashpca/flashpca --bfile filename -v -d 100.

I started this on Monday (so 5 days ago) and it is (according to the verbose output) still reading blocks. Is this within the expectations concerning runtime? 

Thanks for your help!

Tobias 

[Gad Abraham]

Hi Tobias,

Thanks for your message.

Hard to say without the details of your workstation, but two things I
can think of:

1. Do you really need 550k SNPs? Usually I LD prune the SNPs
beforehand with PLINK to something much smaller, say 50k-100k. That's
better for capturing population structure etc (reduces LD).

2. Do you really need 100 PCs? it's quite a lot...

Regards,
Gad

> --
> You received this message because you are subscribed to the Google Groups "flashpca-users" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to flashpca-user...@googlegroups.com.
> To view this discussion on the web, visit https://groups.google.com/d/msgid/flashpca-users/dc9911ee-e530-41f0-86b5-d41b2c2f8e60n%40googlegroups.com.