Tuesday, January 10, 2017

Perseus Nature Methods paper!

EDIT: Honestly, you should probably skip all the stuff I wrote below and check out this recent paper detailing how powerful Perseus is!  (Shoutout to @UCDProteomics for the paper link)

Over the years I feel like I've been just a little harsh on the amazing free software the Max Planck Institute rolls out for us every year. I don't mean to be. I'm honestly a huge fan! Seriously. I only switched over to using commercial proteomics packages (PD 1.2, Mascot, PEAKS) when I felt like they caught up to what I could get out of the release of MaxQuant that I had instructions for.

In the end, I might honestly think that MaxQuant/Andromeda/Perseus is probably better than any commercial software out there. The downsides are linked to one of the big upsides of all open software. It's free. And evolving. The Max Planck software, honestly, is also kind of intimidating. Holy cow...that is a lot of features!!

But...the power....you have is seriously amazing as a user!  This is how I stack it up in my head:

I only made the first part -- and then I felt like it was confusing. Then I tried to clarify it at the bottom. Then some neurons way deep in my brain started making me quote Skeletor (warning..audio)  -- thanks...brain...not sure where our keys are but you're holding onto a 5 minute monologue from a Dolph Lundgren movie...we're gonna need to have a talk about that....what was I talking about?

The top part!  There is no factual basis to that chart. It is just how my...obviously glitchy?...brain stacks these things up.

As a user -- commercial software is super awesome. There are instruction manuals, the software is tested for bugs by people who get paid to test it and there people you can blame if you aren't getting the data you want. You are, however, generally limited in the features you can set and the outputs you can have relative to the other things on the list.

The people at Max Planck, one could argue, are pretty good at proteomics. They design expert software for expert users. It can be intimidating to pick up, but you have a ton of control over the input the output -- the beautiful experimental designs.

The only way you can have more power -- complete power over your input, output, design, everything else -- is to go full out bioinformatician. You break out the R and the Python and you tell the data what the heck you want it to do!

Wow. That was a lot of words. Maybe I should summarize it at the top and here. Perseus is awesome! You should check out this paper!


  1. Surprised you didn't pick up on this in June when it came out - obviously not following your twitter feed as onerously as you should https://twitter.com/PastelBio/status/747555134083862528 and been on the Pastel BioScience resource page since then as well ;-)

  2. It also lacks a user manual anywhere near as good as ProteomeDiscoverer's.