Oxford Nanopore metagenome read length weirdness - TapeStation of the input library shows very dominant peak at 20 kbp, what we actually got from the MinION was mostly 10kbp or lower (mean read length ~5kbp). This didn't happen for pure culture genomes on the same run. Any ideas why, anyone? - ThreadSky

ianpgm.bsky.social • 1 day ago

Oxford Nanopore metagenome read length weirdness - TapeStation of the input library shows very dominant peak at 20 kbp, what we actually got from the MinION was mostly 10kbp or lower (mean read length ~5kbp). This didn't happen for pure culture genomes on the same run. Any ideas why, anyone?

1 / 2

Comments

willaanderson.bsky.social•1 day ago

The accuracy of a rapid electrophoretic DNA separation technique like the tape station at those sizes is fairly low. Look at the order of magnitude that curve covers. There is little chance is separated well.
Meanwhile your nanopore data is a particle-by-particle DNA length distribution analysis.

ianpgm.bsky.social•1 day ago

Thanks! OK this makes sense... not that it makes our lives easier when we're trying to optimise DNA extraction and library prep from tricky samples, but I can see your point.

willaanderson.bsky.social•1 day ago

I hear you. It’s a bit challenging to get real length info for DNA above 10kb using standard electrophoresis techniques. I’d recommended running some low % (0.5) agarose gels at lower voltages (50 V) in TBE for a couple of hours over tape station or lab chip for this for better qualitative data.

willaanderson.bsky.social•1 day ago

I should also mention that electrophoretic techniques report a signal based on dna mass, so smaller fragments will giver a lower signal, even if their molarity is higher, I.e a mass distribution. Nanopore will report a number distribution, so molarity at different sizes is shown.

brooklynkid53bskys.bsky.social•1 day ago

There is a direct solution for this:
Sage Science sells a power supply that lets you run pulse field in a std gel box; you can get good resolution for DNA up to at least 80kb (way beyond the reptation limit)
This is a very simple thing to do, although run times are bit long (~ 5+ hours)

brooklynkid53bskys.bsky.social•1 day ago

Here is the product
https://sagescience.com/products/pippin-pulse/

There are some images at the end of the user manual
https://sagescience.com/wp-content/uploads/2021/06/Pippin-Pulse-User-Manual-460022-Rev-F-1.pdf

willaanderson.bsky.social•1 day ago

Fantastic! Thanks for passing this on. It's exactly what i've been looking for for a looong time. :)

antobeck.bsky.social•1 day ago

That’s nice old Skool!!

willaanderson.bsky.social•1 day ago

I mean, it works "okaay". Someone really needs to make a low cost PFGE for this kind of QC....

antobeck.bsky.social•1 day ago

Sounds like your next project

ianpgm.bsky.social•1 day ago

I understand that the flow cell prefers short fragments, but this seems extreme...

pedalheadphx.bsky.social•1 day ago

If there are strand nicks this could happen, molecule is 20kb, but if a sequenced strand has a nick at 5kb you get a 5kb read. If you think it’s preferential short read, load, run for an hour or so, pull off library, wash and reload, on the promethIOn this often increases N50 significantly

ianpgm.bsky.social•1 day ago

Ah yes, nicks could be likely... this is a challenging sample. Thanks!

Comments

Posting Rules

Reply