Only want to identify proteins (but not quantify them) use any column you want!

 

Proteomics is in such an interesting place right now! There is so much new interest that I hear people talking about proteins in my day-to-day than genes. Maybe that's environment, but it's definitely the first time ever.

Now, you've kind of got two groups of people in proteomics. 

You've got the nitpicky nerds who aren't just satisfied by knowing what proteins are present. These jerks want to know how much of each protein is there. For no reason at all I'm going to call them "Medical doctors and good scientists". Purely, and only, because I randomly referred to them first in this list. 

Then you've got the people who don't care at all how much of each protein is around. They just want to identify a protein is there. It could be one molecule or it could be 90% of the total seven kilograms of protein present. I'm going to call them "Not the other thing", again, I need to come up with two lists with these categories and they had to be in some order. 

Now, the second group generally uses aptamers (SomaScan, Illumina Protein Prep). But if you want to use a mass spectrometer improperly, you can absolutely do this without aptamers and this has been very popular recently. I've seen LCMS data where people are measuring 1-2 scans across peaks. Which might honestly be able to generate "quantitative data" as bad as Illumina's proteomics solution if you hit the peak wrong a couple of times and don't have too many peptides per protein.

This is absolutely not what this new MCP paper is about. 

These authors probably generally fall into the first group, but I think they were just asking the hypothetical question for this second group of scientists "IF you only cared about peptide identification numbers and you didn't care about quantifying proteins, does it matter what chromatography you use?" And, surprisingly, the answer really does appear to be no. This Astral is so fast that any "chromatography" that can get you a 3 second wide (FWHM) peak can get you a whole ton of identified proteins. 

The authors also show that if you run serial dilutions against each other you can see that the lower dilutions seem to have lower protein levels. I'm a little surprised by this because I figured DIA-NN would normalize protein levels by default, but that doesn't appear to be the case. Actually, I just looked and cross-run is enabled by default, so wouldn't that scale the 0.1ng up in intensity to as close as it could get to the 50ng? They probably just turned it off. 

Either way - it's pretty impressive you can pack just about whatever you want into an HPLC column and run 3ng fill time narrow window DIA on an Asstral and identify a whole ton of proteins. I'd love to have seen this study with a spike in experiment to know what you could quantify with each one, but I hope I almost always fall into that first group....