All right. Thank you very much, James, for inviting me to the session today. As I said, my name is Jason Dugas. I work at Denali Therapeutics. I’ve been there for about eight years, and currently the director and staff scientist in charge of the Automated Lab Facility and also Assay Development Group.

Maybe like several of you here, I didn’t actually start in automation. I came from a research background and kind of evolved into automation, which I’ve been doing for about the 10 years or so. And also with my company, I’m working at a much smaller company, so I often have the remit, instead of building one device to do one thing really, really well at high throughput, which we saw several people talking about fantastic builds in that sort of fashion, I instead often build Swiss Army Knives, where I have the funds to get one device that I need to be able to do a lot of things. So instead of doing one thing over and over again, repeatedly, very, very focused, I have devices that can do eight different types of processes over the course of a week with 12 different users coming in and using them for whatever assay they might like to do.

So with that in mind, I was trying to build a protocol here which could adapt onto a Lynx platform which doesn’t have any hardware built into it. So I can take these hardware items on and off for other protocols, whether we sell culture in there or normalizations or serial dilutions, et cetera.

One other shout out I’ll do quickly aside from this, which I was talking to people at Dynamic Devices yesterday. Owning a device where a bunch of users are using it repeatedly, the error handling on the Lynx is the easiest, best system I’ve ever worked with. It’s because it’s the opposite of a Hamilton.

Sorry. What’s that?

Okay, anyways, the fact that users can crash it and come in and you can just throw in a command line and hit step, recover samples, and then they go, “Oh, I’m halfway through. What do I do?” And you can just jump back into the protocol and start over again is fantastic. So thank you for having your software do that.

Okay. So with that said, touchpad. There we go. Okay.

So I was tasked with building a protocol that could at least semi-automate the process of RNA and DNA preps with a wide array of sources, but not having to put any permanent hardware on the deck. So instead of going with some of the MagBead protocols we talked about earlier, I was going with your standard silicon-based protocols from QIAGEN where they have these nice little– and I don’t have movies today, but I did bring props, 96-well plate arrays, which you can load on and work with getting this to be able to automate at least a lot of the liquid handling steps for this protocol. This is the general protocol that QIAGEN has for this kit.

And the initial steps, the lysis, et cetera, because we are working with several groups that have different inputs, be it cells, be it tissue, et cetera, that step we decided to let the users handle on their own because where this all converges is when you end up with those lysis samples in a nice 96 deep well plate. That became easy to automate where we could add the ethanol to those samples, mix those, get them onto the arrays, and then try and get the liquid through these plates all the way through till the final steps.

So before going into the details of that too much, I will have to thank Ben and Scott who worked on several iterations of the final product of this base, which is here, which I’ll go into a little bit more, but thank you so much for all the testing of this back and forth. Working with Dynamic Devices was great and very responsive to my needs to get this final prototype working and very operational in our deck.

So, there we go. So we decided to go with a vacuum-based system. So now we can have vacuum attached to these devices. So all of the liquid addition steps, it just flows through, straight through without having to change anything on the deck.

These bases, which are made, have a nice little silicone seal around here, which means once there’s vacuum pressure applied, this vacuum holds all the way through from the beginning of the protocol, so there’s no manipulation needed there.

These are removable. This is a standard SBS footprint, so this just goes onto your deck in whatever locations the deck maps say. And as you can see there, there’s some tubes coming up, which just goes to a vacuum source underneath the deck. So again, no need to manipulate the deck itself.

A few comments about the vacuum pressure. We initially were excited about the positive pressure aspect of the 96 VVP head, and we tried that initially. However, eventually we decided to go with the negative pressure solution instead with a few advantages. One is we actually get a better flow rate with the vacuum versus the VVP head, even with a separate regulator to get increased pressure out of that head.

But the other issue just was a matter of speed, because with the vacuum pressure going on, and as you can see from the lines there, we can get up to four of these on the deck at once, we can run 384 preps at one time, and the flow is really quick from everything. With the positive pressure head, you have to go and stamp onto each one and push it through. So that just increases the amount of time. Plus, you have to go back then and swap tips out and swap between the adapter plate for pushing air through versus the tips and drop those back and forth. So we really gained a lot of speed as well by just having a constant flow vacuum working with these bases.

On the back of the rig, we have the collection plates, or sorry, the collection bins for one to four different setups on the deck, and then we just have a four to one manifold, which allows us to open or close whichever number of these you have. They’re color coded for ease of use, so the user’s just plugging whichever color tube, doesn’t matter, turn that on. So you have a good vacuum pressure, no matter whether you’re doing one, two, three, or four of these rigs at one time.

Okay. So we start with your chloroform extracted lysates in 96 deep well plates, and at the start of the protocol, the user enters the sample volume, the final elution volume you’re going to use to get your sample out of these plates, and the number of sample plates one through four. And Ben, or somebody at the time, I’ll talk to you later about how to get those nice little input windows, because this is coming up one at a time. I love those nice little multi-windows I only saw for the first time yesterday, so I’m going to figure out how to integrate that into several of my protocols.

Okay, so the setup of the deck is pretty straightforward. The users have the deck map, which they can look at here. You have separate sample tips for each set of samples, which goes up to a common reservoir of ethanol, pre-mixes the samples, and puts them onto the plates with a 2x sample volume. So obviously, there’s a max volume you’re allowed to use for the input. The users know that ahead of time, so that’s fine.

The protocol is flexible enough that if you have a large volume, you can deal with that, and it just splits the volume transfers. So if you put like 800 microliters of input, you’re putting 800 microliters of ethanol, then it’s just going to put two runs through to make sure everything gets sucked through onto the column.

And then, once that’s running, there’s a timer that goes off between each addition, because I’ve timed the system, so I know how long it takes the flow through to go. So it’s a well-behaved set of samples that’ll flow through at an amount of rate. So we give the amount of time to go through with enough dead time to make sure you get good clearance for each step. And then you use just the same set of tips aspirating from your reservoirs from a height so you don’t get cross-contamination. Again, this speeds things up, because you just have tips going back and forth to the reservoir, loading your wash buffer onto one to four plates. Timer goes off, make sure everything’s pushed out, and then goes to the next set of washes and the next set of washes through one RW1 wash and two RPE washes. So, your standard QIAGEN protocol.

Okay, there we go. Now I will note another thing about this protocol that we have found is that, despite getting full clearance of liquid through these plates, positive or negative pressure does not clear all the liquid off the bottom of these plates. Because of this little nub in here, and then there’s a little collar around it, liquid can form a nice little droplet between those, and no amount of pushing will get that liquid out of there. And this was causing a degradation of sample quality because you weren’t losing all of your wash buffer. So, at this point again, semi-automated, the users can take these, actually put them back on the original sample plates because now that’s a waste plate, and do a quick pop spin. It doesn’t even have to be that five-minute spin, it’s just a quick spin to get any little residual off with G-forces. So now you have a fully clean plate.

Then you come back after that quick spin and just add your elution, add if you haven’t already, elution tips, which are smaller volume tips now, because usually people are eluting in 50 to 100 microliters or less. Your elution buffer, which in this case just tends to be water usually. And that applies that to the plates which are now mounted not in this vacuum reservoir, because you don’t want to suck your samples away, but instead your collection plate. And again, for the same reason that the positive pressure was leaving droplets behind, we don’t use positive pressure here. We put them back in the centrifuge. This does do a full five-minute spin now. So you can collect all of your sample out. You don’t leave any behind attached to the reservoir.

I will say that these plates not only work well– Sorry. I just wanted to do that.

There we go. Oh, and again, the elution tips, again, to save time, use the same elution tips across everything because you’re eluting from a height.

We have validated this not only with these QIAGEN manifolds, which can go down to elution volumes of about 50, 40 microliters, but Zymo also makes a version of this where the membrane is much smaller and thinner, so it needs a smaller elution volume, so we can get down to elution volumes of like 15 microliters in those kits. If you have very small samples where you want them to come out at a certain concentration and 50 microliters is far too dilute. So, this base itself is nicely, and I validated this, very cross-compatible with either of those. And I believe these bases are now available from Dynamic Devices. I don’t know if they’re on a website, but you could ask for them and they will custom print them until they actually put them out in their catalog.

And then, as I mentioned, we just bring these over to the centrifuge, spin them down, that takes five minutes. The whole protocol from front to back, once they have those lysates samples and brings them to you, only takes about 15 or 20 minutes. So it’s a very rapid protocol to get up to 384 samples. My users have been very happy with this throughput, comparing to have to do it manually. It’s a huge time savings for them, especially when they have large samples.

One aspect of this is that, if you have very few samples, it’s tough to do this because, obviously, if you have vacuum and most of the wells are empty, you’re not going to get the same vacuum pull. I do have a user next week who’s going to try and use half a plate and just put power from over one half of the plate and see how that works. That’ll be very interesting to see if that allows you to both do less samples with a high degree of quality plus save reagents and use half a plate and the other half a plate. So that’ll be fun to see if that works, so that’ll be results pending.

But when we did this assay, we went ahead and had the users validate the quality. So again, we can do both QIAGEN and the Zymo kits. We can do volume, elution volumes down to about 45, 50 microliters in QIAGEN, down to 20 or 25 microliters in the Zymo kit to get more concentrated material.

When we compared the two kits, they both gave us very high-quality material, good on both gels and on 260/280 reads, comparable to manual preps that they were doing. So the user’s also very happy, not only with the speed, but of the quality of the material we were able to get out. So with this, we were able to add this to our kit of several different assays we can run on the Dynamic Devices without locking it into just doing this type of prep.

Okay. And with that, thank you. And if you have any questions, I’d be happy to answer them.