issue with direction (sign) of infinite regulation

25 views
Skip to first unread message

Christian Schori

unread,
Sep 19, 2022, 5:41:47 AM9/19/22
to MSstats

Dear MSstatsTeam

We've recently observed an issue with MSstats groupComparison function resulting in a wrong direction of inifinite FC depending on ordering of contrast matrix:

I've attached a minimal example with 2 Proteins ("GFP" and "MshA" (=A0A5P4S7Y3)) and 2 Groups ("D" & "F").

According to the log2 Intensities, we're expecting to see an upregulation of GFP in Group D and a  minus INF for MshA compared to Group FP:
  Protein                            D     F
1 sp|P42212|GFP_AEQVI             30.9  24.4
2 tr|A0A5P4S7Y3|A0A5P4S7Y3_VIBPH  NA    31.5

> comp_matrix <- MSstatsContrastMatrix("pairwise", c("D","F"))
       D  F
D vs F 1 -1

> DP_sel <- dataProcess(minimal_abc_sel)
> PW_sel <- groupComparison(contrast.matrix=comp_matrix, data=DP_sel, use_log_file = FALSE)
> PW_sel$ComparisonResult %>% select(Protein, Label, log2FC)
                         Protein  Label   log2FC
1            sp|P42212|GFP_AEQVI D vs F 6.443467
2 tr|A0A5P4S7Y3|A0A5P4S7Y3_VIBPH D vs F     -Inf

So far so good...

But if I repeat the whole analysis with swaped reference and sample group, this is what's happening:
> comp_matrix <- MSstatsContrastMatrix("pairwise", c("F","D"))
       F  D
F vs D 1 -1

> DP_sel <- dataProcess(minimal_abc_sel)
> PW_sel <- groupComparison(contrast.matrix=comp_matrix, data=DP_sel, use_log_file = FALSE)
> PW_sel$ComparisonResult %>% select(Protein, Label, log2FC)
                         Protein  Label    log2FC
1            sp|P42212|GFP_AEQVI F vs D -6.443467
2 tr|A0A5P4S7Y3|A0A5P4S7Y3_VIBPH F vs D      -Inf

Somehow the sign of the infinite regulation doesn't change here, although the regulation of GFP correctly switched from +6 fold up-regulation to -6 fold down-regulation...

I assume this observation has something to do with the default alphabetical ordering of group names in MSstatsContrastMatrix()...

Can you please have a look into this issue?

Thank you already in advance for your help!

Best,
Christian

PS: I've reproduced this observation with MSstats v.4.0.1, v.4.4.1, and v.4.5.1
minimal_sel.rda

Mateusz Staniak

unread,
Sep 19, 2022, 11:09:18 AM9/19/22
to MSstats
Hi,


thank you very much for pointing this out,
(unfortunately) you are correct, and I added candidate fix for this issue to branch "hotfix-infinitelog2fc" of the GitHub repository of MSstats.
I will test it more before merging to main, but it will make it to Bioconductor soon.


Kind regards,
Mateusz
Reply all
Reply to author
Forward
0 new messages