DIA-NN 1.8.2 beta 27 - feedback

[witszymanski]

Hi!
I do not have Twitter any more (cause it sucks), but I wanted to give some short feedback of my test:

On 24 files, the 1.8.2 version was running approx 30% longer, but identified around 5-15% more protein IDs in complex/high intensity samples, and around 50 to 80% more IDs in super low intensity samples.

1.8.1 Settings:

1.8.2 Settings:

I will now do the correlation and other stuff analysis and post here.

[vdemichev]

Hi Witold,

Many thanks for the feedback!

Yes, the speed of analysis might go down on Thermo Orbitrap data (it's been reported before), but should go up on timsTOF.

Best,
Vadim

[witszymanski]

I am comparing these analyses in the meantime, so I will add some graphs and conclusions once in a while:

> nrow(object_1_only)
[1] 171
> nrow(object_2_only)
[1] 230

[witszymanski]

Is this what you meant by "matrix"? :)

[vdemichev]

Yes!

[MichelMoser]

Hi, this sounds very interesting. I am dealing with low intensity and low complexity data (enriched protein toxins) as well but on a timsTOF. Is there any chance of testing DIA-NN 1.8.2 as well? Thx, Michel

[vdemichev]

Yes, it's public https://twitter.com/DemichevLab/status/1705004380557340677

[MichelMoser]

Great, thank you! Any chance this will be pushed to Git Hub soon? I would love to create a docker container out of this to run on a cluster :)

[witszymanski]

So the 1.8.2 finally works now with TimsTof DIA? :D

…

On Mon, Oct 2, 2023 at 12:19 PM MichelMoser ***@***.***> wrote: Great, thank you! Any chance this will be pushed to Git Hub soon? I would love to create a docker container out of this to run on a cluster :) — Reply to this email directly, view it on GitHub <#807 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADZBRVHN2P4AY42N54NAQR3X5KIKVANCNFSM6AAAAAA5C3E334> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[MichelMoser]

yes, just finished with 2 HeLa quality control runs and the stats look amazing!
It ran in half the time of 1.8.1.
But i will have to do some further analyses to say more... :)

[witszymanski]

I am stupid! Of course, it works on TImsTof DIA data. I was jumping thoughts a bit too fast. Reading DiaNN GitHub while reading something on FragPipe at the same time, and there, afaik, the PASEF DIA is still not implemented.