search

hansenlab / minfi

Devel repository for minfi
58 stars 67 forks source link

getSnpBeta returns different values for SNPs after combineArrays #135

Open JennyvD opened 6 years ago

JennyvD commented 6 years ago

Dear developers,

I encountered an issue with RGsets that contain both EPIC and IL450k arrays - created with combineArray(). It seems that getSnpBeta() doesn't work properly on the combined RGset.

I merged EPIC and IL450k RGsets with combineArrays (outType = "IlluminaHumanMethylation450k" ).

I then applied getSnpBeta() on the merged RGset. For the IL450k arrays, the returned SNP beta values for the 59 common SNPs are identical to SNP beta values obtained by applying getSNPBeta on the original 450k RGset (before merging with EPIC arrays using combineArray). But for the EPIC arrays, the SNP beta values obtained from the combined RGset are not the same as SNP beta values obtained from the original EPIC RGset.

Any idea what's happening? Thanks in advance. Jenny

sessionInfo() R version 3.4.1 (2017-06-30) Platform: x86_64-pc-linux-gnu (64-bit) Running under: Debian GNU/Linux 9 (stretch)

other attached packages: [1] IlluminaHumanMethylation450kmanifest_0.4.0 [2] IlluminaHumanMethylationEPICmanifest_0.3.0 [3] minfi_1.24.0
[4] bumphunter_1.20.0
[5] locfit_1.5-9.1
[6] iterators_1.0.9
[7] foreach_1.4.4
[8] Biostrings_2.46.0
[9] XVector_0.18.0
[10] SummarizedExperiment_1.6.5
[11] DelayedArray_0.2.7
[12] matrixStats_0.53.1
[13] Biobase_2.38.0
[14] GenomicRanges_1.28.6
[15] GenomeInfoDb_1.14.0
[16] IRanges_2.10.3
[17] S4Vectors_0.16.0
[18] BiocGenerics_0.24.0

kasperdanielhansen commented 6 years ago

I will look into it.

JennyvD commented 6 years ago

Thank you!

aennecken commented 5 years ago

Dear Jenny, I had the same problem with combineArrays() using R Version 3.4.2 and Minfi Version 1.24.0. When I updated my R and Minfi to the current versions (R Version 3.5 and Minfi Version 1.26.2) the problem disappeared and everything works fine now. I am still curious though what went wrong in the old Minfi Version. Cheers, Anne

kasperdanielhansen commented 5 years ago

The code accidentally switched the A and B loci for SNP probes of type I (or was it II, I don't recall) - only for one of the two types. Basically just an error where A got mapped to B and B to A, because I have code which was c(probenames_A, probenames_B) and should have been reversed.

On Tue, Jul 17, 2018 at 11:00 AM aennecken notifications@github.com wrote:

Dear Jenny, I had the same problem with combineArrays() using R Version 3.4.2 and Minfi Version 1.24.0. When I updated my R and Minfi to the current versions (R Version 3.5 and Minfi Version 1.26.2) the problem disappeared and everything works fine now. I am still curious though what went wrong in the old Minfi Version. Cheers, Anne

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/hansenlab/minfi/issues/135#issuecomment-405613818, or mute the thread https://github.com/notifications/unsubscribe-auth/AEuhnx8gojiXrVCcoqfQwLnMaS8pmmAsks5uHfwHgaJpZM4TRdGk .

aennecken commented 5 years ago

Dear Kasper, thanks a lot for the explanation and the quick reply!