storeylab / alstructure Goto Github PK

When using the function estimate_d(), if there is missing data in the x matrix I get the error from the bug title:
Error in eigen(Wm) : infinite or missing values in 'x'

I know this might not be easy, but SNP matrices frequently contain some % of missing data.
Is there any way to work around the issue in order to perform a d_hat estimation in the presence of missing data?
Thank you in advance!

Downloading GitHub repo storeylab/alstructure@master
0 [main] tar 12172 fstab_read_flags: invalid fstab option - '0'
Error: Does not appear to be an R package (no DESCRIPTION)

Can ALStructure handle non biallelic data?
The matrix used as input in the examples contains 0 (homozygous "a"), 1 (heterozygous) and 2 (homozygous "A"). But what about a tri-allelic SNP? How should this be represented?
I have tried to convert a VCF file with known tri-allelic SNPs to .bed using PLINK to see how the example from ALStructure would handle it (lfa::read.bed()), but it was pointless. PLINK will consider anything different from the above cases as missing data.
I have tested, and it is possible to load a matrix into ALStructure that contains more than 0, 1 and 2 values. But does it make any sense regarding ALStructure's algorithm?
Thank you in advance!