artefact in volcano plot

[Sam Siljee]

Hi all,

Firstly thank you to the team for providing, maintaining, and supporting a fantastic set of tools. I really appreciate your work!

I've recently run into some strange-looking volcano plots, where a lot of data points appear to have a direct relationship, and line up on the plot. I've attached the produced plots.

In addition, I've also noticed some proteins showing infinite or negative infinite fold change with a p value of 0, however I've just filtered these out.

I've tried the MSstats processed data with DESeq2 for differential abundance, and it doesn't produce the same artifact.

Is this a known occurrence?

Thanks!

Sam

[Mateusz Staniak]

Hi,

is it possible to share a subset of data that reproduces the issue with us? Data can be annonymized by changing protein/peptide/condition/bioreplicate labels.

Kind regards,

Mateusz

[Sam Siljee]

Thanks Mateusz,

The raw file is too large to post, but I've attached the processed data and results of the group comparison

The markdown for my code is here:

https://github.com/samsiljee/MSstats_workflow/blob/main/markdowns/MSstats_PD_SS.Rmd

Best regards,

Sam

[Sam Siljee]

Here is a Google Drive link to the raw data and annotations:

https://drive.google.com/drive/folders/15oquztJtTtbPcC5IXcWuygFDwVkwItT6?usp=sharing

Best regards,

Sam

[Sam Siljee]

Hi all,

I've narrowed down the issue, and it's a problem with the dataset I used. A different dataset run with the same script doesn't reproduce the error.

Interestingly, the pattern of artifact produced varies by normalisation method used in `dataProcess`

Sam