staggered RAW files

[KlemensFroehlich]

Hi Vadim,

I recently forgot to destagger my raw files during mzml conversion. I fixed this now and both results seem to be very comparable when it comes to IDs.

Is there a difference for DIA-NN when analyzing 50% overlap staggered windows between: staggered raw files staggered mzml files (direct conversion, no demultiplexing) de-staggered mzml files (demultiplexing activated during mzml conversion)

I am asking because it would be very convenient if I did not have to convert at all even when using staggered DIA.

Best, Klemens

[t-jumel]

Hi all,

do not want to take away anything that Vadim might want to add, but this is something I benchmarked and wanted to share as it is interesting to see.

Not demultiplexing works decent in our case, just a few % coverage lost and a not too strong additional tendency of ratio compression. I still demux every time in normal analysis, 3-5 % higher coverage and better quant. is worth a few hours of computer time in my case. Analyzing not demultiplexed raw or mzML could still be useful for an asap outlier check if one replicate results in a suspiciously small file size etc., instead of using the ultra-fast mode that has a much stronger impact.

Hope it helps, with best wishes, tobi

[vdemichev]

That's cool, thanks for benchmarking!

[KlemensFroehlich]

hi Tobi, cool colors.... somehow reminds me of a certain publication :D Thanks for the insight! This is really cool.

Best Klemens

[tobiasko]

Hi @vdemichev,

do you actually recommend using the demultiplexing option of msconvert (--filter "demultiplex optimization=overlap_only massError=10.0ppm") in case of staggered Thermo .raw files? I am running DIA-NN on Linux, so conversion from Thermo .raw to .mzML is unfortunately not avoidable. Your documentation says that "Acquisition schemes with overlapping windows are supported" and your mzML conversion settings also don't mention any special action for staggered data. Do I get this correctly?

THX, Tobi