[nanos]

The Tris-HCl is used for the elution from Ampure beads. As far as I know the higher pH facilitates the resolubilization of high molecular weight DNA.
You can definitely use pH8, but you might loose more DNA from the long fraction because it does not come off the beads.

[mkdir]

This 48 hr run generated ~ 300 Mb data, is this a usual case?

We tested a 2 hr run, and only generated 29 Mb data. But the protocol said 48 hr should generate 20 Gb. Are there any tricks to improve output yield?

The Y axis of MinKNOW histogram is labeled as "total # reads". Does this "reads" mean base or DNA fragment? We tested a 2-hr run, the MinKNOW histogram peak is more than 10,000,000. But we only got ~ 4000 reads passed.

Best is to fragment with syringe needle (if you have 4-10ug of DNA available per lib prep)
In my case (for my second nanopore run) I had done the following:

1. Diluted my freshly preped DNA with NFW (you can use TE) to total of 400uL,
2. Sheared it by passing through 21G 4' needle attached to 1ml syringe for 20 times
(or use 26G one for 3-5 times),
3. Than performed an Ampure XP wash at 0.45X ratio (+180 uL of beads). Incubate on the rotator for 10 min. Pellet at 60 degrees angle (help gravity and magnetic forse to work in the same direction for ~5min.
4. Wash beads 3 times with 1mL of 70% ethanol.
5. End repair/A-tailing times were set as per NEB recommendation (30 min each instead of 5). All library handling after this step was done with the large bore tips.
6. Intermediate ampure XP washes were done on 0.45X ratio (instead of 1X),
7. The first steps of the ligation had been done for 20 min on the RT and for 30 min in the fridge (ligating adapter with a huge helicase attached to it to a long (25-50kb) DNA fragment requires cool and quiet conditions for quite a while).
8. The tether ligation was done for 20 min at the RT.
9. Dynabeads streptavidine M280 beads were washed with the ONT kit's bead wash solution, leave to bind for 10 min.
10. Pellet at angle ~60 degrees (to have gravity and magnetic field working in the same direction).
11. Resuspend with extreme care - like cosmid library ligation tube flicking.
12. Pellet again at an angle, and repeat the wash. Elute as per protocol.
13. Make sure to degas the fuel mix and water used to prime the flowcell (otherwise there would be a lot of tiny bubbles over the ASIC array after a few hours @34 degrees).
I had used an Eppendorf Concentrator Plus set to V-AQ mode for 90-180 sec @ 20 degrees.
14. Entire library had been loaded.
15. The run had started at -185 mV bias voltage, and by the 48h the voltage had gone to -250mV.

The modal RAW read length had been 8.5kb, and there was a long tail with a good fraction of 20kb, 30kb, even quite a few 60kb raw reads. Total RAW events had been ~450M.

PS: it looks like the manufacturers adapter ligation conditions had been optimised for the previous version of the kit, when there was no motor protein attached to the adapter during a ligation step. The attachment of the motor changes the adapter's molecular dynamics quite a bit (by increasing it's mass by ~2 orders of magnitude), which increases the time required for efficient ligation to the long DNA fragment - have a read of the BAC/YAC library construction protocols, ligation conditions section, to give you some more background info.

So the results had been a raw yiels of ~446M events, with the following raw reads distribution (see attachment):
M160713_48h_raw_reads_histogram.png

For metrichor basecalling results see attachments: M160713_metrichor_48h_1D_res_c2.png and M160713_metrichor_48h_2D_res_c2.png

[gringer]

This 48 hr run generated ~ 300 Mb data, is this a usual case?

That depends on what flow cell you're running. If it's an R9 flow cell, then 300Mb of basecalled data is reasonable, although with good sample prep and good, fresh flow cells, it can get to a couple of Gb.

We tested a 2 hr run, and only generated 29 Mb data. But the protocol said 48 hr should generate 20 Gb. Are there any tricks to improve output yield?

What protocol? Clive Brown has recently achieved an in-lab yield of 10Gb basecalled with 1D ligation kit + R9.4 flow cell (i.e. the most recent and fastest available kits and flow cells). It would be weird for any protocol to suggest a yield over twice that.

Without knowing anything more about your sample prep, it's difficult to say. If you're going for maximum yield, fragment FFPE-repaired (or PCR amplified) dsDNA into pieces of about 3-8kbp, and make sure you're loading library onto the sequencer that has a hairpin-containing dsDNA concentration (as quantified by a fluorescence-based method) of at least 4 ng/μl. Also, choose a flow cell that has >450 active channels at initial QC, and doesn't have SPDS (sudden pore die-off syndrome).

The Y axis of MinKNOW histogram is labeled as "total # reads". Does this "reads" mean base or DNA fragment? We tested a 2-hr run, the MinKNOW histogram peak is more than 10,000,000. But we only got ~ 4000 reads passed.

It's actually the predicted number of called bases. If there are 1000 reads with a read length of 8kb, then the histogram value would be 8,000,000.

[mkdir]

Hi gringer, thank you very much for the information. The nanopore website suggested 21Gb (Product specifications comparison, https://nanoporetech.com/products#comparison). But you have answered all my questions. Thanks a lot.

That depends on what flow cell you're running. If it's an R9 flow cell, then 300Mb of basecalled data is reasonable, although with good sample prep and good, fresh flow cells, it can get to a couple of Gb.



What protocol? Clive Brown has recently achieved an in-lab yield of 10Gb basecalled with 1D ligation kit + R9.4 flow cell (i.e. the most recent and fastest available kits and flow cells). It would be weird for any protocol to suggest a yield over twice that.

Without knowing anything more about your sample prep, it's difficult to say. If you're going for maximum yield, fragment FFPE-repaired (or PCR amplified) dsDNA into pieces of about 3-8kbp, and make sure you're loading library onto the sequencer that has a hairpin-containing dsDNA concentration (as quantified by a fluorescence-based method) of at least 4 ng/μl. Also, choose a flow cell that has >450 active channels at initial QC, and doesn't have SPDS (sudden pore die-off syndrome).



It's actually the predicted number of called bases. If there are 1000 reads with a read length of 8kb, then the histogram value would be 8,000,000.

[gringer]

Wow! I wasn't aware there were marketing people at ONT who were more optimistic than Clive Brown. Those numbers are technically true, but Clive has said that the current flow cell performance is only about 10% efficient (a bit more for in-lab flow cell testing).

[mkdir]

Hi gringer, thank you very much!

Wow! I wasn't aware there were marketing people at ONT who were more optimistic than Clive Brown. Those numbers are technically true, but Clive has said that the current flow cell performance is only about 10% efficient (a bit more for in-lab flow cell testing).

[apredeus]

Best is to fragment with syringe needle (if you have 4-10ug of DNA available per lib prep)
In my case (for my second nanopore run) I had done the following:

1. Diluted my freshly preped DNA with NFW (you can use TE) to total of 400uL,
2. Sheared it by passing through 21G 4' needle attached to 1ml syringe for 20 times
(or use 26G one for 3-5 times),
3. Than performed an Ampure XP wash at 0.45X ratio (+180 uL of beads). Incubate on the rotator for 10 min. Pellet at 60 degrees angle (help gravity and magnetic forse to work in the same direction for ~5min.
4. Wash beads 3 times with 1mL of 70% ethanol.
5. End repair/A-tailing times were set as per NEB recommendation (30 min each instead of 5). All library handling after this step was done with the large bore tips.
6. Intermediate ampure XP washes were done on 0.45X ratio (instead of 1X),
7. The first steps of the ligation had been done for 20 min on the RT and for 30 min in the fridge (ligating adapter with a huge helicase attached to it to a long (25-50kb) DNA fragment requires cool and quiet conditions for quite a while).
8. The tether ligation was done for 20 min at the RT.
9. Dynabeads streptavidine M280 beads were washed with the ONT kit's bead wash solution, leave to bind for 10 min.
10. Pellet at angle ~60 degrees (to have gravity and magnetic field working in the same direction).
11. Resuspend with extreme care - like cosmid library ligation tube flicking.
12. Pellet again at an angle, and repeat the wash. Elute as per protocol.
13. Make sure to degas the fuel mix and water used to prime the flowcell (otherwise there would be a lot of tiny bubbles over the ASIC array after a few hours @34 degrees).
I had used an Eppendorf Concentrator Plus set to V-AQ mode for 90-180 sec @ 20 degrees.
14. Entire library had been loaded.
15. The run had started at -185 mV bias voltage, and by the 48h the voltage had gone to -250mV.

The modal RAW read length had been 8.5kb, and there was a long tail with a good fraction of 20kb, 30kb, even quite a few 60kb raw reads. Total RAW events had been ~450M.

PS: it looks like the manufacturers adapter ligation conditions had been optimised for the previous version of the kit, when there was no motor protein attached to the adapter during a ligation step. The attachment of the motor changes the adapter's molecular dynamics quite a bit (by increasing it's mass by ~2 orders of magnitude), which increases the time required for efficient ligation to the long DNA fragment - have a read of the BAC/YAC library construction protocols, ligation conditions section, to give you some more background info.

So the results had been a raw yiels of ~446M events, with the following raw reads distribution (see attachment):
M160713_48h_raw_reads_histogram.png

For metrichor basecalling results see attachments: M160713_metrichor_48h_1D_res_c2.png and M160713_metrichor_48h_2D_res_c2.png

This is for 2D protocol, right? Since I figured you don't need Streptavidin beads for the 1D.

Do you have the detailed protocol for 1D? Would/did you change anything for the new flowcell (R9.4)? And if your goal is to get several bacterial genomes assembled, would you shear still or try to get larger fragment DNA? There are some preprints which adjust experimental conditions in order to try and get most fragments above 10k, but it seems to be lowering the throughput of the sequencer.

[Markiyan]

Here is my first succefull attempt at 1D SQK-RAD001 FLO-MIN105 (R9)

The previous method was for 2D library with needle fragmentation.

If doing 1D tagmentation library prep (RAD-SQK001):

1. Use old school DNA prep methods and make sure to follow the lysis carefully, quality is much more important than quantity!
Use wide bore tips for all lysate/DNA handling steps!

2. Lysozyme lysis was followed by with phenol/cloroform (twice), followed by two chloroform washes (to remove all phenol).
3. Precipitate the DNA by the isopropanol and spool it onto glass stick/pippette (flame sterilized with hook at the end) in 15 or 50ml falcon. (you can spin down the unspooled DNA, wash it with ethanol, and use it for other experiments, but it would be less pure, than spooled fraction).
4. Dink the glass stick into 70% ethanol twice (10 minutes each wash in 15 ml falcon tune).
5. Elute into 0.5-1ml of sterile water or 2mM edta elution buffer.
6. Start tagmentation prep as soon, as the DNA is quantified, freeze the unused DNA.

7. For tagmentation, I've took 7.5 ul of ~50 ng/ul DNA (~0.4ug) with wide bore 200ul tip and used 2.5 ul of frm mix for 2.5 mins @30degrees. (further optimisation of the tagmentation conditions, volumes and ammounts of DNA would be needed).
8. Than 1 minute at 75 degrees water bath.
9. let it cool down to RT.
10. add 1ul of RAD and 0.2 ul of ligase mix, incubate for 5-10min @RT, than store in fridge/ice until ready to load.

This method had worked well with FLO-MIN105 (R9) flowcells - yielded 0.9Gevents, including some mappable 100kbp reads but FLO-MIN106 (R9.4) are much more sensitive to osmotic imbalance and didn't work well with SQK-RAD001 kit (without loading beads) provided with them (end of October 2016).

[apredeus]

Excellent, thank you for the insight, I appreciate all the info.

So if the method doesn't work well for R9.4, what would you recommend? A different kit?

[Markiyan]

Use SQK-RAD002 with Loading Beads

Excellent, thank you for the insight, I appreciate all the info.

So if the method doesn't work well for R9.4, what would you recommend? A different kit?

Use SQK-RAD002 kit with Loading Beads, all other tricks are in force. Tagmentation conditions optimisation would be needed.

PS: If SQK-RAD002 work well with R9.4, please can you post the following:
1. Tagmentation conditions: Volume, DNA concentration, FRM mix volume used, tagmentation time.
2. Loading volumes used.
3. Draft reads length distribution chart (from minknow) after first 12-24h of the run.

This would help us to find optimal tagmentation conditions, depending on target reads length wanted.

[apredeus]

Thank you, as soon as we do the run I'll definitely do that.

[morning latte]

Hello,

What vendor do you use for Tris-HCl (pH 8.5)? I know Fisher has Tris-HCl with pH 8.0 but can't find any with pH 8.5.

[nanos]

Your friends are Trizma powder, water, HCl and a well calibrated ph-meter ;-)

[morning latte]

I have one more question to ask.

Do you think it would be okay to use Axygen™ AxyPrep Mag™ PCR Clean-up Kits ($142) instead of Agencourt AMPure XP beads ($385)? I am trying to reduce budget for the first run of MinION.

[apredeus]

any of the experienced users could help? I've used a standard RNAse-proteinase digestion method for our bacteria, but it seems like it was too sturdy, and RNAse got destroyed before most cells were digested. Thus I got DNA with a whole lot of rRNA and needed to purify it once again, losing the valuable material.

What's a good way to avoid it and still get Nanopore-suitable reads?

And, in general, what's your favorite way to extract DNA for 1D?