search

mpc-bioinformatics / McQuaC

Transform the Quality Control workflow from Knime into a workflow in Nextflow
Other
2 stars 0 forks source link

Ionmaps not aggregated #69

Open julianu opened 4 months ago

julianu commented 4 months ago

Would it not be more useful to have the "ion maps" more like conventional RT - m/z maps, instead of the density graphs? I assume, the common user would be more used to these.

KarinSchork commented 4 months ago

What is the z-dimension in a conventional RT/mz-map? I think we just replicated some R-code that Stefan gave us

julianu commented 4 months ago

The z-dimension is the intensity. As this is arbitrary, it is not really fixed.

KarinSchork commented 4 months ago

aaah I see yes, that makes sense. I will look into that

KarinSchork commented 3 months ago

ionmap_test

Do you mean like this? As far as I can see this only uses MS2-data at the moment.

julianu commented 3 months ago

Yes, something like that. But these are only the mz/RT points for each MS2 event, or what does it depict? What you basically need for the ionmap is a representation of the MS1 event per RT, so should be no big problem to create.

KarinSchork commented 3 months ago

Ok, I will look into that with @Luxxii , because currently not all information needed is exported for MS1

julianu commented 3 months ago

Well, basically you just need to go through each MS1 and plot the values as a segment in the RT range and put all these together. That's all the information needed. Ho did you create the other plots? You basically need the same info for these, and could also use them and set the density to 0 or something.

KarinSchork commented 3 months ago

The former plots just count how many MS2 were measured in a specific m/z and RT window. So it doesn't use the intensity.

Yes, we would need to export the MS1 information in the result table then, because at the end I only have this table and not the intermediate files (and I think we should leave it like this because in the future I could then plot older runs by pulling the data from the database).

julianu commented 3 months ago

Ah, ok, I thought these plots were the plots I described, as I only knew them under the name ionmap. If the plot with MS2 is also useful, we should probably give more accurate names.

KarinSchork commented 3 months ago

Ok, I see, I will ask Stefan for a better name (because he suggested it based on an R script)

KarinSchork commented 1 month ago

image

This would be an example with the current values we have in the data (ms1_map_rt_array, ms1_map_mz_array, ms1_map_intens_array). I reduce the datapoints to 1Mio, otherwise the plot takes ages to generate (the example QExHF file has around 16Mio data points). This is also not interactive, I would save them as pngs then.

Does this look okay? Any suggestions for improvement?

KarinSchork commented 1 month ago

I added these plots in commit 19e7bb8