It looks like genome enthusiasts have got a nice Valentine’s Days gift from Oxford Nanopore today. Oxford Nanopore have started announcing the lucky participants who will be receiving their Oxford MinION as part of the Oxford Nanopore’s early access program. It is not clear, how many labs were given the early access to MinION.
I’ve heard of four people getting MinION access programme emails so far today! Woop!
— Nick Loman (@pathogenomenick) February 14, 2014
Not just that, two years after first announcing their products, we got a first look at Oxford Nanopore data at today’s talk by David Jaffe at AGBT 2014. Here is a real quick summary of what do we know about Oxford Nanopore data from the tweets storified.
The idea sequencing by nanopore has been around for about two decades. Now we have a product from Oxford Nanopore to do that. Every one know how small Nanopore’s MinIon looks like.
David Jaffe presented data from two samples one on E.coli samples and the other on Scardovia wiggsiae, a gram-positive bacillus found in oral cavities with genome size of 1.6mb. As one would expect, the sequencing experiment was done by Oxford nanopore, not David Jaffe’s group. As told by Oxford Nanopore, the basic steps of library prep involved starting with
high molecular weight DNA to have enough DNA to sequencing, fragmenting the DNA without size selection, and ligating adapters.
One end of the double stranded DNA to be sequenced is tethered to make the sequencing process easier. And ratcheting enzyme sits on the nanopore to make it single stranded DNA pass through the nanopore. As the single stranded DNA passes through the nanopore, the nanopore system measures the DNA sequence interms of current change.
What Do We Know About Oxford Nanopore Data?
The current that is produced is function of number of bases present in the nanopore, not a function of a single base. And this makes the whole sequencing process and inferring the seqeunced base difficult. We can only wish getting signal at single base level resolution! It looks like, in the current setup, Oxford Nanopore is measuring signal of 6-mers. The raw data from MinIon is a profile of current (or voltage) as a function of time. David Jaffe (and Oxford Nanopore?) called the data as Squiggles.
The computational challenge the Oxford Nanopore scientists are dealing with is converting the squiggles from k-mers into sequence data. The approach is proabilistic in nature ( one can imagine some kind HMM is at the work here). Once inferring the k-mers from the squiggles, Oxford Nanopore constructs a read from these k-mers. So the overall sequencing process is like this: Pore -> signal -> kmers -> bases. And the presented data was obtained with 25bp/s.
The MinION sequences a full DNA fragment that is fed to, with no real size limits. In the data presented by David Jaffe, median read length is about 5 kb with some reads as long as 20 kb. And the bioanalyzer trace of DNA sizes nicely matched the read lengths from MinION.
What Do We Know About Errors in Oxford Nanopore Reads?
What about the read errors. The reads do have errors and some do look like systematic. In the data showed by David Jaffe, reads had no error for a stretch of bases and then finds a chunk of region ‘error-blocks’ with error prone bases. There was also systematic deletions; for example 5 out of six reads had a missing T. David Jaffe (or Oxford Nanopore) presented an interesting error metric and said that “84% of the reads have at lease one perfect 50-mer and 100% have of the reads have at least one perfect 25-mers”