minION from oxford nanopore, is it still so promising?

Question

[Buffo]

Years ago we participate in the pre-launching program of minION, however its low performance (to us) we decide to leave the program, personally I had very good expectancies about it, and further up-dates looks very promising, is someone actually using it? what about the actual sequencing performance? what about the new flow-cells versions?

*In this particular case, I prefer personal opinions (if possible) than published papers,

[GenoMax]

I prefer personal opinions (if possible) than published papers

You are willing to trust strangers than a (peer-reviewed) publication :-)

[Buffo]

nice try genomax XD

*In this particular case, I prefer personal opinions (if possible) than published papers,

[Buffo]

and also, what about Biostars? This is not a journal XD XD XD XD XD XD

Answer 1

[Asaf]

We used it through an outsource service for bacterial genomes. The reads were better than I expected, combined with Illumina reads we got almost a whole chromosome. The problem with the system (as we saw it) is that the costs are a bit obscure, you can use the flow-cell several times but you should wash it after each run (with expensive reagents) and it's not clear how many times you can use each flow cell and how its age impacts the results.

I just saw somewhere that they sequenced RNA directly with good results, I don't remember where.

[Buffo]

Yes, on my attemps we assembly about 9 Mb (total contigs) from a genome expected to be around 50 Mb after 4 or 5 sequencing runs :(, very bad performance meanwhile from pacbio we obtain a complete chromosome (1.3 Mb) using just one flowcell.

[Asaf]

They had a real bump in accuracy after May 2016. In addition I only used 2D reads, the median quality was above Q20. It gave long contigs but they were still full of errors, usually in poly-nucleotides (i.e. is AAAAA, AAAA or AAAAAA?), combined with Illumina reads the assembly was really good.

Answer 2

[igor]

From limited personal experience, each new flow cell is a significant improvement over the previous generation. If you last used Nanopore more than a year ago, you will be surprised by the difference.

I just saw a good blog post on this topic (as it relates to PacBio, which is really the most relevant context):

MinION can deliver similar data yields as PacBio. The read lengths can be longer -- much, much longer -- but the data quality is decidedly poorer at the consensus level. Despite the many improvements in ONT's base calling -- documented spectacularly well in Ryan Wick's preprint README file -- Oxford's consensus sequences hit a wall around 0.2% error rate but PacBio can get much, much better with sufficient depth.