Question about Cosine score and cosine similarity

[Ting Yi Li]

Hello everyone,

I just found that there is a different score between Cosine and Cosine similarity in GNPS results.

I want to ask which one should I refer to when I need to know the annotation quality compared to reference data.

Thanks a lot.

[longkang cao]

Hello, have you found the answer to your question?

[Ángel Sánchez Illana]

Hello, If I am not wrong, I believe the issue lies in the fact that the score employed in GNPS is the modified cosine (as detailed in the GNPS papers, utilizing the shift to do a kind of neutral loss alignment), and also applying the filters you've chosen (by default, to a 50 Da window and the precursor filter). Not sure if also using only the matched peaks.  In contrast, the cosine score in the USI is without filters and can be normal cosine or the shifted version as used in GNPS. Some expert in GNPS could confirm my hypothesis? 

[Mingxun Wang]

Yes, that is correct I believe in my reading of your message. However, depending upon where you are generating the USI, the query spectrum is also filtered. 

Best,

Ming

--
You received this message because you are subscribed to the Google Groups "GNPS Discussion Forum and Bug Reports" group.
To unsubscribe from this group and stop receiving emails from it, send an email to molecular_networking_b...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/molecular_networking_bug_reports/45dfaa00-36b8-42b9-8ced-8fa2e714690bn%40googlegroups.com.