Wednesday, June 17, 2015

Reproduce the awesome new Nature mass-tolerance search with Proteome Discoverer and Byonic

For about 6 hours my PC has sounded like a jet engine is taking off.  Why?  Because I HAVE to see if I can replicate the findings in the Nature paper from Chick et. al., with Proteome Discoverer.

Here is how I set it up.

Is it perfect?  Of course not.  But its a start.  Yes, you have to have Byonic for this, or at least something with a "delta M" or in this case "wildcard search" option.

Now...I did have my Byonic node in PD 2.0 set to "Heavy" CPU usage.  When I got back from visiting a customer this afternoon the  processor was pegged at 100% usage.  I'd have put a screenshot in but I didn't have enough processing power to actually take one (and I appear to have misplaced my phone....somebody want to call me?)  So I crashed the whole thing out and went into Byonic settings and changed them to custom and decided to only use 6 of my physical CPU cores.  This gave me the usage picture above (all 6 cores are pegged out!)

I also realized that I'd never get to the end of the search today with the full Uniprot database.  (And I actually have to do work with this PC at some point! Collaborators are waiting for data!) so I created a custom database that only contains human kinases.

The data file in question is a 200ng injection of the Pierce HeLa digest ran on a 2 hour gradient on somebody's Q Exactive Plus instrument (separated on a 25cm column, I think).

As suggested by Joel Chick in the comments of the last blog entry I dropped the number of missed cleavages (in this case, I decided to use 0, again, I've got to do some work with this thing!)

So...How'd it do?

Not terrible.  It knocked the search out in less than 1.5 hours.  It came back with a total of 875 PSMs.

I extracted the PSM data to Excel and then removed anything that was a Cys +57.  This left me with 169 wild card PSM matches.  If I follow the paper, I would line these up by mass and bin them (histogram Excel 365 function, FTW!) but I'm too lazy for that.

I was pretty much just going to answer this question:  can I perform a +500Da mass-tolerant search with Proteome Discoverer 2.0?  Yes, if I have Byonic.  Will it be painful with a full organism FASTA?  Probably, but it IS definitely possible.  The second question is, did I get new stuff?  You bet your sweet peppy!

1 comment:

  1. Hi Ben,

    Byonic seems to be really useful for this. But how does it handle multiple neutral losses?