Skip to main content
Topic: Very, very weird quant issue in MS-DIAL v4.92 (Read 299 times) previous topic - next topic

Very, very weird quant issue in MS-DIAL v4.92

Documented here: https://x.com/drchrispook/status/1795768077893923092

A random subset of peaks in my data are being underidentified in MS-DIAL processing (v4.9.221218). This only shows up when I run a PCA and get a massive division in the data. It's about 10% of features and the difference varies feature-by-feature. The one I'm presenting here is cholesterol as it's a good example, but there are many others. Most of the features show no such issue, however. My internal standards are fine.

I'll attach images of the quant to another post.

 

Re: Very, very weird quant issue in MS-DIAL v4.92

Reply #1
Agh1 is the MS-DIAL quant. Look at the difference between the large (2e7) and small peaks (2e6). When I look at them in TOPPView (agh2)  or my own Python integration script that uses PyOpenMS (agh3) they're all at 2e7. There's no difference in mass accuracy between any of the runs. It's <1.5ppm. This is very weird and I don't like it. I've never seen anything like this in >5 years of MS-DIAL analysis.