chrchang / eigensoft Goto Github PK

Hi

I am trying to PCA with smartpca for 200K sample , 100K SNPS with 256G memory .
However, the process does not finish (core dumped). I suspect memory shortage, but I do not have access to additional memories.

Is this problem is reasonable(compared with memory size)? and Are there any way to solve it?

Thanks,

Satoshi

Hi,
I often work with draft assemblies that consist of hundreds of contigs. Because of that, I have needed to change the parameter "MAXCH" in the file "include/mcio.h" to something much bigger. (It is "100" by default.)
I don't understand why such a limit is needed and why has it been set so low. Could it be removed or set so high (e.g. 9999) that it wouldn't break the analyses (at least the format conversion) of contig assemblies?

Best,
Ari

I ran 'EIGENSTRAT/example.perl' and find no new plot files created, although 'EIGENSTRAT/example.perl' contatin the plot params.

I directly run ./ploteig under 'bin' and warning raised -bash: ./ploteig: /usr/local/bin/perl: bad interpreter: No such file or directory. Then I tried add perl in front of the commond and plot files occured.

I am not sure it is due to my environment or the software.

I ran 'EIGENSTRAT/example.perl' and find no new plot files created, although 'EIGENSTRAT/example.perl' contatin the plot params.

I directly run ./ploteig under 'bin' and warning raised -bash: ./ploteig: /usr/local/bin/perl: bad interpreter: No such file or directory. Then I tried add perl in front of the commond and plot files occured.

I am not sure it is due to my environment or the software.

Readme say's to contact this e-mail but I'm getting a bounce:

The following addresses had permanent fatal errors -----
[email protected]
(reason: 550 5.1.1 [email protected]: Recipient address rejected: User unknown in virtual alias table)

My question to you was:

Base R's prcomp() command returns a number of components along with eigenvalues (for example: sdev, rotation, center, scale, x). Is it possible to get such components out of SMARTPCA?

According to the documentation (https://github.com/chrchang/eigensoft/tree/master/POPGEN) it looks like the only outputs you get are evecoutname and evaloutnames (eigenvectors & eigenvalues) so I'm not sure if this possible or if you know of a workaround?

Thank you for your time

Hi,
I have been getting this issue even if I change the ids.
idnames too long 100_0301001891 100_0301001891 ll: 41 limit: 39
./smart_pca_without_hapmap.sh: line 114: 1717458 Aborted (core dumped) /EIGENSOFT/4.2/bin/smartpca

Please help me find and rectify the error.
Thanks in advance.

Regards
Mano

Suppose we have the following files:
in.bed, in.bim, in.fam

The command:

smartpca.perl -i in.bed -a in.bim -b in.fam -s 6 \
-e out.eval -l out.elog -o out.pca -p out.plot

results in out.pca having only zero entries. On the other hand, using plink to recode the binary files into *.ped and *.map, the commands:

smartpca.perl -i in.ped -a in.map -b in.ped -s 6 \
-e out.eval -l out.elog -o out.pca -p out.plot

and

smartpca.perl -i in.ped -a in.map -b in.fam -s 6 \
-e out.eval -l out.elog -o out.pca -p out.plot

both compute the same eigenvalues as the previous one, but the PCA scores are correctly computed.