Interview: PromethION 24 and human disease research
Fri 18th June 2021
with Eric Cabannes, Development Laboratory Manager at the Centre National de Recherche en Génomique Humaine (CNRGH)
Conducted and written by Jonathan Pugh
"We now routinely obtain between 100 and 160 gigabases of data for DNA...over 30x sequencing depth for a human genome"
discussing data output from a single PromethION flow cell. PromethION 24 can run up to 24 flow cells, and PromethION 48 can run up to 48 flow cells
In December 2020 at the Nanopore Community meeting, VP of Platform Technology at Oxford Nanopore James Clarke referenced a number of cumulative production improvements for the PromethION flow cell. “We’ve come up with a new protocol…that gives us about 20-25% more nanopores”, he said. “Anyone that runs these platforms knows that more nanopores really means more data output”. In my discussion with Eric Cabannes in April 2021 it was clear this, in addition to improved consistency, is something he agrees with; “I've noticed since around October, November 2020 the quality of flow cell has greatly improved… this to me is the biggest improvement…the quality, the robustness, [the] reproducibility of flow cell”.
Eric is the Development Laboratory Manager at CNRGH, the French national research centre using high-throughput sequencing to enable the optimisation of genetic and genomic research on human diseases. He oversees the introduction of new technologies into their production environment with a focus on single cell sequencing, long-fragment whole-genome sequencing (WGS), and full-length RNA sequencing. Since October 2018 they have been using PromethION and are now running after having been upgraded “very recently, not even a month ago on to the PromethION 24. The move has been very smooth”. This device is capable of running up to 24 PromethION Flow Cells, while a version for 48 flow cells (PromethION 48) is also available. Samples for sequencing are provided by collaborators on a range of diseases from neurogenetic, metabolic, developmental and ophthalmic. Often they are related to rare diseases where investigations are focused on trying to find and characterize genetic determinants, such as structural variations in the genome, which may explain the pathological conditions.
In order to facilitate the generation of long reads for WGS they have become experts in high molecular weight DNA extraction. “We’ve been practicing the Nanobind CBB Big DNA Kit from Circulomics almost since the beginning” he was proud to say “and that was obviously a great help for this long read technology”. Their WGS work on PromethION focuses on both structural variant identification and genome de novo assembly. For these applications they sometimes shear into the 10-20kb range, sometimes they leave the DNA unfragmented and are obtaining reads “up to the megabase size range”.
The importance of extraction
The expertise they have built up makes them confident in how to get the best from the technology. “We spent a lot of the time at the beginning on the purity and quality of the material and this is essential with such technology. We discovered very early the Circulomics kits and a few other providers such as Revolugen, now we've a solution to extract high molecular weight DNA”. When I asked Eric what other changes they had made to help optimise their long-read sequencing, he responded simply “this is the one, if you know if you have great material”. He summed it up well: “A nanopore can sequence a megabase, providing you can extract a fragment that size”. His sentiments match those of Oxford Nanopore’s Vânia Costa; “you cannot sequence what you don’t have”. When it comes to long reads and good results in sequencing - you need a great extraction method.
"A nanopore can sequence a megabase, providing you can extract a fragment that size"
To elaborate on this, Eric detailed a problem they encountered on one of their collaborative projects “We had a problem sequencing DNA provided by collaborators; the classical QCs were okay (purity and quantity as read by nanodrop and qubit, respectively), everything seemed to be fine”. Despite this they were seeing an unexplained reduction in throughput after the first day of sequencing, while an identical project being run at the same time was not experiencing any issues. With Eric’s expertise, the extraction was his major concern. To resolve the issue once and for all they went straight to the source. “We finally extracted DNA from the blood specimen of this patient for this program, and it worked very well”. A lack of control over sample extraction can be problematic for centres such as CNRGH, but Oxford Nanopore is consistently adding to its extraction protocol library to help with education in this area.
Interactions during the run can also help improve the quantity of data obtained. “To increase the yield we've done the classical refueling” Eric informs me, referring to the addition of fresh ATP containing buffers during sequencing to maintain the speed at which the enzyme ratchets DNA through the pore. However once their current project is complete “what we should do, what is [in] the plan and we will do shortly, is [to use] the new release kit - the SQK-LSK110. This has been developed to avoid refuelling and we will try it soon”. Ultimately however his message is simple “There's not much to improve; new material is, I would say, why your flow cell will succeed”.
His team at CNRGH are well setup to run their experiments on PromethION now, with no single expert required. “There was one person that first used the PromethION and then transfer to others, and it worked fine…everyone in the lab is independent and autonomous to work, to run the technology that's not a problem”. Multiple experiments can also be conducted in parallel. “We run up to eight flow cells simultaneously” he replied when I asked him how they used the device’s flexibility “[with] eight flow cells we had an average of 120, 130 Gigabases per flow cell”. However, this did come with a data burden: “we store the data locally on the PromethION server before being transferred afterwards and stored on our servers… there is a limit in the storage capacity of the hard drive…we had to transfer this data and then erase on the local drives”. Imminent new updates in MinKNOW will help to ensure data can be offloaded smoothly in real-time, rather than relying on rsync transfer, and improved default compression formats also ensure space is used more efficiently
At the end of our discussion I asked Eric what he thought the future of nanopore sequencing held. “Oxford [Nanopore’s] technology I definitely think is a very elegant way to sequence” he responded, “I'm following what is still in Research and Development - the solid state nanopore… and I'm sure that people in the protein field are waiting for long long time already, to…be able to sequence peptides and protein, very rapidly”. In the future perhaps Eric and his team, among others, will be adopting nanopore peptide sequencing for their research into human diseases.
Jonathan Pugh is an Associate Director at Oxford Nanopore Technologies and has spent almost 10 years developing and introducing nanopore sequencing technology
Want to learn more?