Okay. So I was going to start with another Hamlet thing "to use long gradients or not to..." or something else equally stupid. So, I obviously first Googled "Pug Hamlet" for the 100th time in my life. This time I got a good image! And why? Because some crazy person (not me, I swear) has a KickStarter up where he's trying to put on an entire production of Hamlet just using pugs as actors. I love when people do stuff that makes me feel more sane. You can find it here. And yes, it got funded.
There was a point to this rambling. Recently, a paper out of the Max Planck institute, who I consider big fans of long columns + ultralong gradients, showed that with a QE HF, moving from a 2 hour to 4 hour gradient did not increase peptide numbers significantly.
Now, I have a number of friends out there in the field that are using instruments not quite as screamingly fast as the QE HF who are getting huge ID numbers using big columns (btw, I consider big columns 35cm on up) and I'll be visiting a great lab in a few weeks that uses 100cm columns for most experiments and it works really well.
In an interesting analysis in JPR, Hong Wang et al., systematically optimize ultra long columns and gradients on their instrumentation. Ultra long? Yeah, they essentially run a 110 cm column before their liquid ESI junction followed by another 50 cm column. So, 160 cm of total separation power.
I'd almost say we have a controversy here, but we really don't. If you are sequencing at 20Hz or higher, you are going to get to the bottom of your identifiable/ionizable/detectable/fragmentable peptide list pretty fast and you may not need out of this world chromatography to do it. If you are running slower than this, you are just going to need to go to extra lengths to get the same or better coverage.
Definitely an interesting paper. The consensus between the two, interestingly, is that we can get some darned nice coverage these days with new instrumentation and no pre-fractionation!
longer columns are not always better. for example, in the second paper you've mentioned, they use 5um particles. Van deemter equation shows that the minimal plate height of 5um will always be higher than 1.7um particles, due to eddy diffusion coefficient differences. Similarly, longer columns suffer from dispersion problem, and require an increase of flow rate, which results, in turn, in penalty in mass transfer.
ReplyDeleteOverall, i'd say the super-long column may not outperform a pepmap 50cm 2um columns - quite the contrary.
David,
ReplyDeleteThanks so much for the comments. These are important points to take into account for sure. We focus so much on the evolution of the instrumentation and data processing side that we sometimes forget that the chromatography resins have been evolving as well. In a related note, I've seen cases first-hand where simply moving from a name brand C-18 that is a few years old to something with the exact same name on to box that is brand new has resulted in improved peptide IDs. Our thought was not that the original particles had gone bad, but that the procedure for producing this product had improved in some way over time to produce better particles. (Though there is no way of ruling out that it just wasn't as good anymore.
Again, thanks for highlighting such an important factor!