This is the first image that showed up when I looked up POSTNOVO. Google translate was somewhat less than exceptionally helpful, so would someone let me know if it says something reeeeeaaally bad? Thanks!
Big shoutout to Hugo (from Porto!!) for reassuring me that I didn't post anything bad above. (An old friend of mine has some large Kanji symbols tattooed on his arm in college. He was a victim of a somewhat hilarious prank and they don't mean what he was told they mean. To make it even funnier, he lives in Japan now, with an awesome joke tattooed on himself forever. You can never be too careful!)
And I looked up POSTNOVO because of this Just Accepted Manuscript at JPR!
I'll be honest. I forgot my work computer and I can't actually read this. But the abstract says it's like Percolator for de novo searches.
What's the biggest problem with de novo? FDR!!!
Whats Percolator best at? Fixing FDR problems when it has loads of data to look at. De novo produces loads of data (mostly bad, and I'm not being mean) and its really hard for normal engines to look through so much. When PepNovo came out (that's Ari's one, right? I'm pretty sure) I set it once to allow up to 20 results per MS/MS spectra. And regretted it because a LOT of them came back with 20 possible sequences. How do you search through all that without going crazy, becoming nocturnal, and adopting a bunch of elderly dogs? No idea.
But you could go and hunt down PostNovo's Github (here) and dig around.
What do you find? You find that it doesn't look all that hard at all to connect this awesome new tool to the DENOVO GUI, the ultrapowerful software from the amazing people at CompOmics.
And if that isn't enough for you to look for your laptop and read the paper, probably nothing is!
Clarification: This isn't Percolator, but it's described as post processing for de novo, like Percolator is for regular searches. I wasn't clear.
No comments:
Post a Comment