hijacking this thread to have some collective brain work
My setup is on a MagEx star and the labware I use are square well Axygen 10ml plates. (I do an extraction in 8mL primary sample). After all the previous clean-ups there are still residual liquids left in the 4 plates I use. But the residual is not equal. We have different residual volumes left over in each plate. Teaching is correct and we use the Alpaqua 24FLX magnet (with springs). I aspirate all supernatant (1mL) from the last step with 1000uL tips and after a short shake (600 rpm) the plate is put back on the magnet for 30 seconds. After that I aspirate residual liquid. For some plates we remove all liquid, for some there is as little as 3uL left, for some 20uL.
I remove with the 96MPH so I can’t do a touch off on the bottom like you would do with a individual channel.
I came a cross this resource that summarizes some of the issues in optimizing the bead extraction protocols for automation: Automating Nucleic Acid Extraction (promega.com) (Video)
I don’t necessarily agree with everything but find it to be a very good summery of the challenges .
Just if anyone is interested
I’m interpreting this as running 4 plates simultaneously with 4 magnets, 4 HHS, etc… My first steps troubleshooting that would be:
check if z_bottom of magnet 1 is truly the same as magnets 2-3. Possibly pause the system while during the liquid waste removal step and physically pick up the plate to gauge distance. Apply individual offsets if there are inconsistencies
Manually observe the sample wells after primary liquid waste removal to get a good understanding of where this residual vol is coming from. The beads themselves could just be clumped differently leading to different performance. If you can find a time point that consistently shows residual volume has settled, add your 2nd aspirate prior to drying to ~30s longer than that timepoint.
Oversaspirate supernatant volume to ensure that the excess volume isn’t a cascading effect from previous waste removals. Usually 1.1*expected volume is more than enough to get residual + extra.
If all else fails, add a prompt to visually inspect. A working product now is better than a perfect product later.
We’ve just made some progress on the original issue recently which might be relevant to @Pascal enquery too. In our case, in line with the advices above, longer elution time with intensive mixing by pipetting to resuspend beads made the biggest difference in terms of improving DNA recovery. I also increased binding time from 5 min to 10 min, added additional aspiration step to remove residual ethanol, and drying beads for 3 min before the elution. Elution itself is for 5 min with 50 pipetting cycles.
I dropped heat-shaker completely because it didn’t seem to homogenize samples particularly well and I opted for pipet mixing. That being said , I work with the regular PCR plates, not deep wells, and so limited in the applicable rpm range for the heat-shaker. But in anycase, my feeling is that mixing by pipetting may circumvent some of the issues.
all plates are taught individually for their Z-axis. The X are all the same (and used in sorting, so I would opt for X to be the same). Regarding the offset, great idea!
2)I aspirate the first 1000 uL from the supernatant in two steps. Step 1 is 950ul and step 2 is 50. After this I dispense in the liquid waste and remove tips.
I perform a residual removal with 300ul tips, in this case I remove at 1mm from bottom 75ul and 0.7mm from the bottom again 25ul.
Strange thing is that visually, I see good results. But not across the whole plate.
for the aspiration from the Alpaqua, is the teaching such that the Zmax height actually compresses the spring down by 1-2 mm ?
the typical 2-3 mm travel of the spring should be taken advantage of,
I’ve found that, for optimum residual liquid removal, you perform the bulk of the liquid removal is you’ve described for 1 mL tips
for the final 300 uL tip removal, aspirate slowly with the spring/plate compressed slightly and retract slowly - this way, the internal tip suction can aspirate any liquid remaining - even better is to handle as maybe 3-4 cycled aspirations of 20-30 uL after 5-10 second delays to allow liquid to settle to bottom of the well
i’ve also seen occurences where the magnets do not perform equally - sometimes the spring mechanism
Sounds like you’re doing everything right, so maybe just a strange hard-to-nail problem?
It could help to aspirate the liquid waste twice for large volumes - aspirate ~1/2 volume at bottom+0.7mm, wait 60s, aspirate 2nd volume. This could expedite in magnetization as well, or potentially improve yield.
What is your bead volume? This could just be wash solution stuck in your bead mass after washes. A good way to check is to pick up your plate after supernatant removal and swirl it (like a fine wine) to resuspend the beads/wash solution. This makes it easier to see and guesstimate the volume remaining. If it’s irregular across a single plate, I’d blame 96MPH alignment - possible XY plane rotation.
If troubleshooting just isn’t bearing fruit, utilize the springs and aspirate at z_bottom+0.00mm and increase aspiration volumes to full volume followed up by the 50/25µL aspirate to ensure ~everything~ is removed. Alpaqua claims that the springs prevent vacuuming - I never truly trust that, but Olaf overengineered the hell out of those things from the original beckman design.
@Optimize brings up a GREAT point! Check for rust or sticking on those springs.
Just want to bring something up that I’ve experienced first hand but if you have strong magnets in successive sites/nests, they CAN create a relatively strong magnetic field.
Why does this matter? You can potentially see this manifest when you push down on the spring and all four edges (or just one corner) doesn’t immediately spring back up. Ergo you can have edges or sites that aren’t at their true height because the magnet interferes with its ability to spring back into to place which will lead to variability in your removal steps.
Also want to chime in and repeat that if you’re not using a 3.0mm orbital shake with a round bottom plate or KF style plate, you will never achieve exceptional efficient mixing without also having to pipette mix for high volumes.
The residual liquid originated from a different spot in the whole assay. I was babysitting my workflow yesterday and noticed that the JSB liquid had a tendency to foam (even looking at it made it foam). The foam/bubbles caused my volume not to be 100% and therefore the first removal was there with some foam. On average this foam was 30uL and this was exactly what was left over.
For now I am not removing 100% of the first sup but added a 2nd removal after 30 seconds. No more residual liquid left. Awaiting QC results to see what’s going on.
Elution is done on the shaker. I do not see anything wrong with this part. But for this setup the shaking speed is 1500 rpm with a accelleration rate of 750
I’m going to assume that is Omega Bio-tek’s JSB Buffer - probably for cfDNA extraction by looking at your sample volume.
I used to work for Omega and - while I didn’t develop the LCs for that buffer - I did work with it frequently and it is a very tricky buffer. Shoot a message to their support line and the automation guys there can help you get this working in no time. They could probably even send you a LC to save some time unless the policy has changed.
You are correct! It’s the Omega cfDNA extraction kit.
Supernatant is now removed correctly by adjusting a few variables for the primary removal. Still it’s a tricky liquid to remove, but it get washed away proper by GT7 and SPW.
We had a lot of trouble getting our DNA extraction with magnetic beads to match manual (using Promega DNA IQ kit on hamilton starlet). Our situation is different since we target small amounts of short DNA… But a few key things for us were:
1 - Removing plate transfer steps (everything is processed in original plate)
2 - Dealing with foam misleading the cLLD and not pipetting correct volumes
3 - Handling a precipitate issue (our input sample had a reagent which reacted with the promega DNA IQ reagents, causing a precipitate which the DNA got caught in and was disgarded as waste) - we changed the input reagent to fix this. Note the precipitate only happened if that reagent was above a certain amount.
4 - Temperature (a temperature setting was incorrectly set at the start vs at the desired step)
I know many labs have had good success with double elution. e.g. if you want a final 200ul of DNA, you do the final elution in 100ul , then go back to the beads and elute again with another 100ul. This greatly increased their yields (+30%)
