What is the biggest problem with de novo peptide sequencing? I think the consensus is definitely false discoveries. There are some clever solutions out there. PEAKS, in specific, has worked very hard over the last several years to develop and improve their FDR calculations, and to make those mechanisms public.
This new paper (currently in press at MCP, so open access) from Leprevost et. al., has taken a good hard swing at this problem and I really like their approach.
This paper introduces us to PepExplorer, a new algorithm that can compile the results of different de novo engines and compare them to help us get a feel for the confidence of our findings. For example, if PepNovo+ and PEAKS (which are completely different algorithms) both give you high scores for a sequence identity (and, even better, it makes it through PEAKS FDR!) then we can be a lot more confident about that result.
I like the approach, but I can't speak to the software yet. I'm pretty busy leading up to this ASMS thing. If you happen to check it out, please post your opinions in the comments section. Hopefully I'll get around to putting up my impressions later in the summer.
You can download PepExplorer here.