Curious what everyone’s approach is to sample mixing when it comes to reducing the risk of x-contam. I work in a high-specificity NGS lab where I’ve recently removed all mixing using orbital shakers for fear of cross contamination. This seems overly-cautious and not data driven, but I’m willing to learn more and change my opinion.
I can’t imagine this being higher than background noise unless there are multiple rounds of PCR with massive overamplification, but I’d be interested to hear what the general consensus is on when to use orbital shaking as a tool and when it is too risky to employ. I’ve done previous experiments on cross contamination where I’ve seen trace levels (barely above NTC) but has anyone else done studies on this?
We run two weekly QC plates through our whole process to check for any contamination, both cross contamination of desired amplicons, and of template into NTC control wells. The QC plates are a checkerboard pattern of two different amplicons and we have empty wells in the middle of those plates that are our NTC controls.
We don’t have a need for orbital shaking. My biggest concern was with the XPeel after PCR. My thought was that would throw amplicons all over the place. It hasn’t proven to introduce any contamination at all. But, looking back post-PCR would require quite a lot of contamination of one amplicon into another well. So it’s unlikely that we would see anything. Still, we don’t see any template contamination into our NTC wells, and that template plate also has an XPeel deseal step.
What’s your need for orbital shaking? We just do pipette mixing for any mix steps. It isn’t always the most efficient because we then do have to take every plate to a liquid handler and use tips, but so far this step is batched into other steps that also require the liquid handler anyway.
If i have two HHS I can mix two plates at once but I only have one 96MPH.
I actually ran a test this morning where I put 3 µL Trypan blue inside a 96PCR well near the tip-touch height, above 50µL of water below in the well. I had to shake at 1500rpm with 1.5mm radius before I saw one of the 3 droplet reps fall and mix with the sample. I know its not conclusive but It was a test I could run in like 15min which gave me some confidence that if there were contamination it would be minimal, like you mentioned.
I’ve only recently been in an NGS lab but this seems like a common concern from the MT’s for potential contamination - is this just an old unproven assumption?
Basically, any time you would mix/vortex your sample in a manual workflow, you will have to use an orbital shaker in automation. You probably won’t reach the necessary mixing depth by using just tips, so there isn’t really an alternative. However, considering the low volumes and that all mixing steps are normally done in midi or deep well plates, the possibility of cross contamination is neglible.
There’s definitely a lot of superstition around NGS, and PCR in general. That’s why we decided to run QC plates, now we have actual data. Because invariably if there’s a problem the blame will come back on the automation. This way I have actual proof every week that the process is working as intended.
The contamination that I most worried about is actually prior to PCR1 in an NGS workflow. After PCR1 your product is going to outnumber any other contamination by many, many orders of magnitude. After PCR2 everything is barcoded and can be pooled. For PCR2 you do need to be concerned about your barcodes contaminating other barcodes though.
I think I would recommend pushing the boundaries. Setup a test experiment if you can and make it as dirty as possible and have control wells that should be blank. Are they still blank when you run it?
the general consensus is on when to use orbital shaking as a tool and when it is too risky to employ
I don’t know if there is a broad consensus because it’s so context specific. As @jnecr notes there are definitely parts of the workflow where it’s a good idea to exercise caution but there’s also ways to engineer around those concerns as @PatrickM points out with the labware.
I always frame my work around the intention and context. If my intention and context align, I’ll do tip mixing. However I tend to default to orbital shaking where the context and intention make sense. Since my labware and orbital shaker are part of the context. I design my workflows to bias the intention and context.
If my intention is to agitate the container materials, I probably do need a high orbital RPM for a 2.0mm rotation and maybe a deeper container well in the labware. If my intention is to mix, an orbital shaker of 3.0mm allows me to mix more thoroughly without using high RPM, I just need to determine if that’s 1000RPM at 30 seconds or 500RPM for like 75 seconds.
If I pipette mix, the intention also matters so I look at speeds, heights, tracking and try to standardize that as well with evidence. A homogenization mix vs pellet resuspension are going to be different.
With that said, there’s also assay design choices that can impact the aforementioned. I’ve seen some plate maps that are truly horrifying in my life.
FWIW, in my own testing using standard 96PCR plates (the ~180uL ones) I’ve only truly seen issues when the rcf gets above ~20xg. At around that mark, you can start to see droplets near the top of wells jostle and break free - however there likely won’t be cross-talk between wells without droplets present at or near the well openings.
At around 25rcf is when you get full, catastrophic, worktable-sanitization level events where sample is flying outside of wells with some impressive distance. The difference in 5 xg will depend though, since the relationship is [orbital radius]^2
Here’s my shitty 15-minute data from testing. Happy 50th birthday to Excel, btw.
It’s crazy to assume that there is any significant aerosol generation or other X-contam source when the forces being generated here are equivalent to someone holding the plate at arms-length and spinning moderately fast, around 1 rotation per 4.5 seconds.
Re: Data - I’m missing some context, is this a visual check?
Re: Final paragraph - Are you saying that orbital shaking doesn’t generate significant forces, so worrying about cross-contamination from it is overblown?
It seems like you’re pointing out a false relation, but I think its safe to assume that the assumption can be “true-ish”. When looking at something more relative (such as rcf) and comparing the forces generated from shaking on the orbital to something like a normal plate spindown which would be closer to 500rcf you get an apples-to-apples dataset. Also adding the comparison to someone literally spinning in a circle is a nice visual for anyone who isn’t scientific or familiar with orbital shaking - which is kind of what this post is about since it appears from these responses that the cross contamination risk is, for the most part, a misconception in most applications.
Don’t get me wrong. I love the visual images of dye flying all over a plate. But, if you’re doing NGS you have one of the most sensitive platforms for detecting cross contamination available in science: PCR.
Setup an entire plate with master mix and primers but no template. Throw some random template in a few wells, some template droplets up near the edge of the wells and give it a good shake. Run it through 35-40 cycles of PCR and then do a quick gel. You’ll find out real quick if that template is getting spread into any other wells.
100% agree, the dye was just a quick 15 minute test! My MTs are very visually oriented so the blue dye actually had great impact when I shared the pictures. The pcr test is next, time permitting, but our template is expensive and custom synthesized so I’ll need to order some supplies before I can generate real data.
Well if you template is that specific perhaps contamination of template can’t even be possible. In my case the template is from various human cell lines. Being that I am human, and even most of the people I work with are human, the presents a lot of opportunity for template contamination.
I’m curious about the claim that a powerful enough orbital shaker can generate aerosolization - though at a high enough speed the force might become similar to sonication.
In my experience, any liquids with reduced surface tension (compared to water) have issues in an orbital shaker as it will creep out of the well. I’d avoid any shaking, even with a strong plate seal, if cross contamination is a concern.
One quick-and-dirty experiment I have done is to tape a piece of SpotOn to the top of the labware with a representative amount of liquid in it, and then test orbital shaking patterns, RPMs, and durations. If droplets are flying up and out of the wells, they appear as bright blue spots on the paper. This is not nearly as sensitive or representative as the PCR-barcode test that @jnecr described, but it is fast, cheap, and easy. Thank to Brandon Kwan-Leong for teaching me that one.
I recall seeing somewhere (maybe here?) a paper that is failing to turn up with some quick googling that showed worse cross contamination in square wells than round or those with rounded corners. I think this was with creepy crawlies rather than NGS but I’d imagine the physical mechanism is still there…
