MBR and smart profiling

[xiaosiruwei]

Hi,Vadim,

I saw your answer in discussion # 443, which mentioned “smart profiling - only extracts medium- and high- quality empirical spectra, while for precursors with low quality spectra the original library spectrum is retained.”

Can I ask a few questions here?

1. In MBR mode, a new library will be generated from DIA analysis. If it is in smart profiling, the first step will be through Global.Q.Value to choose precursors?

2. As for the extraction of empirical spectra, could you please explain a bit more about how to determine the medium, high, and low quality of empirical spectra? I am curious about this.

3. Additionally, before executing the Second pass, I did not see the initialization of the new library to generate decoy in the log? I think decoy generation is necessary before analyzing sample data.

Best,
Weixiao

[vdemichev]

Hi Weixiao,

1. Yes, in any of the profiling modes.
2. This is quite a complex and not yet published algorithm.
3. Decoys are always generated on the fly in all cases.

But in general, I would suggest to use IDs, RT and IM profiling, this seems to perform best throughout all datasets we have seen so far.

Best,
Vadim