WARNING: it is strongly recommended to first generate an in silico-predicted library in a separate pipeline step and then use it to process the raw data, now without activating FASTA digest

[jeewan-rijal]

Hello, thanks a lot for this excellent tool and regular updates.
In the latest versions of DIA-NN, we see this warning as shown in the title.
Does it apply to all datasets? Do we now need to run samples in 2 steps?
Or can we go from the fasta to the final result as we did before?
What difference should we expect?

[vdemichev]

Hi,

Yes, please always heed DIA-NN's warnings and see if they make sense. If it says 'strongly recommended', then please definitely address the warning. I have just posted also an explanation for the reasons for this particular warning: #1238. That being said, in 1.9.2 'on-the-fly' one-step approach should work fine.

Best,
Vadim