Discrepancy between GUI options and R code output

[Yasmin Chau]

I tried out the GUI to generate the starting base code for running MSstats in a terminal (so I can use my university's cluster computational resources). I think there is discrepancy between the options I select at the data upload page and the default options that are selected in the R code that is generated at the end. 

Especially for: filter_with_Qvalue (default is TRUE in the R code but on GUI it is unchecked), removeProtein_with1Feature (default seems to be TRUE in the R code, but it's unchecked in the GUI).

I attached screenshots of my GUI screen to show those options, and I attached the resulting R code. 

[Anthony Wu]

Hi,

May I ask which version of MSstatsShiny are you using?  I believe this issue was fixed in the most recent release (Release 3.23, MSstatsShiny 1.14.0, R 4.6.0).  If you're using an older version, could you update to the latest version and let me know if you still see any discrepancies?

Thanks,

Tony

[Yasmin Chau]

Hi Tony,

I am indeed using the latest release, release 3.23 with MSstats Shiny 1.14.0 and with R 4.6.0. The results I shared came from those releases. 

Either way, for my current use I can change the settings as I like once I run it on the terminal. Are the default settings in the manual still the recommended default settings to use?

Best,

Yasmin

[Anthony Wu]

Oh my mistake, it looks like we had fixed the discrepancy a week after the release 3.23, so those changes aren't available to users yet.  

In terms of my recommendations, I usually enable q-value filtering and removing a protein with a singular feature because those peptides/proteins are typically noisy and unreliable.   Removing low quality peptides/proteins can indirectly help with downstream statistical analysis because it can lead to higher quality quantifications, and it can avoid an unnecessary increase in the multiple hypothesis testing burden from noisy hypotheses that are unlikely to yield significance.

Tony