search

andrewparkermorgan / argyle

An R package for import, QC and analysis of Illumina Infinium genotyping arrays
31 stars 10 forks source link

Option suggestion #2

Open jcgrenier opened 7 years ago

jcgrenier commented 7 years ago

Hello @andrewparkermorgan,

I just found your tool, it seems really great and pretty clean! I might have a suggestion, since the problem is presently occurring for me. I'm trying to do imputation analysis of some data that I got back from a long time ago and have still some trouble figuring which version of the array that has been used at that time.

Do you happen to have any solution to find this? Once we know the correct version we can then apply some scripts coming from the tool given by the McCarthy Group to put back our array so we can compare it with actual human reference releases.

http://www.well.ox.ac.uk/~wrayner/strand/index.html

Thanks for your help!

JC

andrewparkermorgan commented 7 years ago

At this time argyle requires that the chromosome, position and alleles be known a priori.

There is a utility for guessing alleles from observed genotypes, and updating the marker map accordingly -- see ?fix.strand.swaps. When merging two genotypes objects we also provide an option to attempt reconciliation of allele order (A1 vs A2, ie. ALT vs REF) and strand (ie. A/C vs T/G). See ?merge.genotypes and option check.alleles. Those will only work for sites at which the minor-allele count is greater than zero.

But argyle does not offer any utility for guessing which genotyping platform was used for a given dataset. Unless you have access to the raw, unfiltered data, I would consider it very hazardous to try to infer genotyping platform and then apply imputation. There are so many ways for that to go wrong.