I Don't Know How Long A Revio Run Takes...
I don’t know how long a Revio takes to run. I suspect most of my readers are not so interested in “seeing my working”. So here are my best guesses on the time taken during each stage of the sequencing process from “sample-to-sequence”:
The sample prep times comes from PacBio’s documentation for the Sequel IIe. I suspect this all stays the same for the Revio. My uncertainty comes from 3 steps which on past systems have taken between 0 and >12 hours… my current best guess is 35 hours from sample to sequence based on the estimates below.
Trying to figure this out…
The first part of the process seems to be relatively understandable… you’ve got all this stuff:
Which you need to do to prepare your SMRTBells… and can be summarized as follows (discussed here):
After this you still need to perform the “ABC” step. This includes primer and polymerase binding… here for the Sequel II they’re saying 15 mins. But this used to vary by sample type. On top of this you’ve got the on instrument “adaptive loading time” and “pre-extension” time.
On older kits these varied a lot by sample type. Polymerase binding time varying between 1 and 4 hours, and pre-extension time between 0 and 8 hours:
The protocol for the Revio Polymerase Binding kit doesn’t seem to be available… but I think it’s likely similar to the Sequel II 3.2 binding kit… what binding time should you use here? You need special software to tell you…
If I use the waybackmachine to find the newest version of this document that actually lists times here’s what we find:
Which suggests that polymerase binding times and pre-extension times have been trending downward. My guess is therefore that polymerase binding is currently ~1h (perhaps less) and pre-extension is ~2 hours.
But…
What the hell is pre-extension anyway?
I didn’t really understand what pre-extension was… well I understood what it was… but not why you need to do it. This required hunting down this publication,1 which describes the process. Essentially, many polymerases/templates just break due to template damage. With pre-extension you overload the wells, so that some have multiple templates. Then you run the polymerase for a while and let the ones that are likely to break, break. This it seems increases your overall yield of good polymerases/templates.
Which bring us to the leading causes of terminating reads on PacBio’s platform:
Pre-extension should help with the first two, leading only disassociation and laser damage. It would be interesting to know which of these dominates and if solutions here could help improve performance further.
What About Adaptive Loading?
And then you’ve got the on instrument adaptive loading time. How long does that take? No idea… but it seems you can set a 2 hour limit on this in software. I’m pretty much left guessing here. Given the 2 hour limit, throwing in an estimate of 2 hour doesn’t seem unreasonable.
TLDR
My best guess at Revio sample-to-sequence run time is 35 hours. I don’t know how long pre-extension takes. But I at least know what it is now! We have also confirmed that reads are likely limited by disassociation of the polymerase and laser damage!
If you have any better estimates please do reach out (new@sgenomics.org) or on the Discord.
Relevant quote:
Thanks for this analyses Nava, getting estimate numbers on the TAT sure is complex!
"It depends on the application" indeed... and that 10h total time (per your flowchart) doesn't include quant, nor sample prep pre-genomic DNA QC (just as an FYI as I'm thinking through this). How much additional time would you estimate these to take? Let's assume say 24 samples for WGS on the Revio (which is about the throughput, they claim 1300 30x WGS / year or 26 per week if a year is 50 weeks).
Perhaps an additional 3h for 24 samples? (I'm assuming something simple for quant like a Qubit, but am not sure how sensitive a Revio or Sequel is to overloading...)