Wednesday, February 07, 2024

VoltaLabs Launches Callisto for DNA Extraction & Library Prep

Here at AGBT, VoltaLabs has unveiled their 24-sample DNA extraction and NGS library prep Callisto instrument, which is particularly suited for long read applications but is also suited for short read work. Volta has matured liquid handling automation to a novel open top electrowetting technology. Priced at $125K and planning to ship in the second quarter, Callisto is designed as a walk-away solution requiring no human interaction during a run. Personally, not only do I love the a new medium-throughput instrument for HMW DNA extraction and manipulation, but I also can at least pretend I helped steer the company In that directions

Injecting Myself in the Story

I was first introduced to Volta CEO Udayan Umapathi in the Fall of 2018 through a Warp Drive Bio colleague who knew him and thought we might talk productively. . Volta had a small space in The Engine, an MIT-backed incubator space in the heart of Cambridge. Udayan asked me what applications I saw as unmet needs in NGS, and so of course I said that HMW DNA extraction was a highly labor-intensive process and similarly constructing long read libraries by ligation. 

Once I landed at Ginkgo in 2019, I reached out to Udayan again. After an extended chain of emails, we met face-to-face again at AGBT 2020. Then followed more intense discussions as the pandemic raged, as we expressing our strong interest in accessing early Volta instruments for HMW DNA extraction and processing.

By Spring 2022, Volta was ready to show their progress in a very confident manner - by inviting myself and an expert lab colleague to visit them at there expanded space in The Engine and to watch a 2-sample prototype in operation. Correction, to not only watch but for my colleague to actually load a one sample instrument - and for us to sequence the resulting material. So we made a library from Volta-purified HMW DNA and we had sheared material we brought over made into a library at Volta and in parallel made into a library at Ginkgo. On sequencing, both pairings looked like twins - there was no overt difference between the Oxford Nanopore library made with Volta-extracted DNA vs. manually-extracted DNA from the same source material, nor was it easy to differentiate the PacBio HiFi library made on Volta vs one made manually from the same sheared DNA.

Last year at AGBT the two of us met again with Volta to discuss the Minimum Viable Product - what instrument specs, documentation requirements, consumable ability and so forth. We said eight samples was what we hoped for; Volta said they were aiming higher.         Sadly l, my colleague was poached by another biotech soon after - well, sad for me - it was a stellar opportunity  and only a fool wouldn’t have taken it.

Electrowetting Library Prep History

Each of the 24 Callisto modules is based on electrowetting, a technology in which a patterned surface (called VoltaNode) can have its hydrophilicity changed by changes in applied voltage. The volume of liquid moved are larger than many microfluidic technologies, but a key attraction for HMW operations is very lot shearing forces.  Electrowetting has been attended before forDNA extraction and library prep, with decidedly mixed results

NuGEN’s Mondrian was the first try, but never achieved strong market penetration. Mondrian could process eight samples per run. NuGEN emphasized a wide range of Illumina library types and that their technology could enable library construction from as little as 1ng input - though this was really just a case that they kits could support that and the Mondrian didn’t degrade the performance of the kits. But customers did not like the oil which was required - could be messy to pipe the and the cartridges could leak oil.

NuGEN had licensed the electrowetting technology from Advanced Liquid Logic, and so was blindsided when Illumina bought ALL.  So long Mondrian.  Illumina launched Neoprep at 2015 AGBT. A 16 sample instrument - but many of the same issues persisted:  such as leaky cartridges spilling oil.  So two years later Illumina pulls the plug on the product.

Oxford Nanopore was the next company to try electrowetting, and was confident they could solve the problems.  Their VolTRAX unit was portable, much like the MinION sequencer, and inexpensive.  But there were many years delay from initial announcement to launch, more revisions after that and it was never easy to find committed users.  Newer versions could even run PCR onboard and an ARTIC PCR for SARS-CoV-2 was released for it - but due to the limited number of sample ports, VolTRAX could never run many samples at a time.  And the oil leaked.  This past year, ONT quietly removed VolTRAX from their product pages.  They are still promoting the concept of “TraxION” which would marry MinION sequencing to a VolTRAX unit as an integrated sample-to-sequence solution, but we’ll see if they can pull it off.  And if the oil stops leaking.

Miroculus brought out the Miro Canvas device at AGBT 2020, just as the pandemic was entering the public conscious.  A single sample instrument with the ability for users to design their own protocols (oft promised with the other instruments, unclear if ever delivered).  Still uses oil, though I haven’t heard a lot of complaints about it.  A number of groups developed protocols for long read extraction and long read library prep - both excellent fits for an instrument with very low shear forces. Miroculous was acquired by Integra last March - it made sense for a product that had entered technical maintenance phase to be folded into a large automation vendor with established sales and support channels.


Enter VoltaLabs Callisto

Okay, with the sad history of Mondrian, NeoPrep and VolTRAX instruments flopping with electrowetting technology, why am I excited with VoltaLabs Callisto?  What’s differentiates them from all these other instruments.  Great question, and there are multiple answers.  The first, simple one: no oil.

Callisto represents an even greater design difference than all the others, a differentiation Umapathi emphasized during that first meeting (well, he also emphasized no oil!) - Volta has an open surface design.  The VoltaNode cartridges are a simple plastic frame which holds four electrowetting surfaces, as illustrated below

Because the design is open, all of the surface can be easily accessed by a pipette – one driven by a human arm or better on an automated liquid handler. Every previous instrument has had a limited number of sample ports. In the final instrument, only a robot will access the VoltaNode surfaces, but it must have made development much faster to not be forced to work through fixed ports with a long design cycle to move the size, location or number of ports.

The electrowetting surface also doesn’t suffer from one limitation of some prior systems, in that each square within it can be used many times, whereas others had limits on the number of times they could receive a liquid.  Since liquids were required to enter those systems via a limited number of ports, programming a protocol required careful tracking and management of the squares close to the ports.  Volta eliminates that issue on both fronts: the squares aren’t limited and the liquid handler can deposit or retrieve liquids from wherever makes sense.  

Having an integrated liquid handler and full access to the electrowetting surface offers some interesting opportunities for protocol optimization.  When I visited recently, one scientist was running through a new protocol and I watched as while the magnetic system was swirling around a mixture, the liquid handler was pre-positioning reagents for the next step on other portions of the surface. By the way, it’s really, really geeky cool to watch the magnetic bead system work with HMW DNA, as the beads line up and then you see these chorus lines of black beads swirling around - a Busby Berkeley spectacle for the molecular biology set.

Each Volta module also has thermal control and magnetic bead manipulation.  A protocol with multiple master mixes/buffers, multiple incubation temperatures and many bead cleanups is the ideal fit for Callisto - and that’s exactly how one can describe the HiFi library prep for PacBio - and Callisto is already ideal due to the low shear forces. In the image above, you can see a progression across the four panels from a droplet containing beads, a magnetically-focused set of beads within the droplet and then the last two panels showing the beads being removed from the liquid.

Volta already has a suite of protocols implemented on Callisto and has more in development.  The first protocols will include DNA extraction for short and long reads, PacBio HiFi library generation, a nuclease-shearing ligation library protocol for short reads, a protocol automating portions of a short read hybrid capture and a simple DNA bead cleanup. The pattern for all of these is a relatively simple bit of hands-on time to place the desired number of consumables in the machine – as the image shows above they are in cassettes of 4 VoltaNodes, with up to 6 such cassettes for the maximum capacity of 24 samples per run.  A set of deck positions in the front accept tips, reagent tubes and sample tubes.  The instrument then runs hands-off until the final DNA sample or library is ready at the end. In industry in particular, freeing an employee from hours of hands on time is highly valued. This hands-on time savings can be very substantial; one person producing a dozen HiFi libraries manually can spend all day at the bench; with Callisto this becomes a half hour of instrument set-up and a much simpler run cleanup. With the times of the protocols, one person can easily prepare 24 HMW DNA via VoltaPure in the morning, have a leisurely lunch, and then prepare 24 HIFi libraries using VoltaPrep in the afternoon for loading before leaving work - and during both Volta runs they are free to be productive

For most protocols, Volta plans to sell repackaged versions of well known library kit brands, organized for easy loading on the Callisto.  The company claims that due to the miniaturization of the reactions, Callisto will be cost-competitive with running the kits manually.  Some kits, notably the PacBio HiFi library prep, will come with the Callisto-specific reagents and customers will add to the carriers the PacBio-sourced reagent tubes.

At an AGBT breakfast presentation, the Broad's Brendan Blumensteil emphasized the attractive aspects of Callisto to their operations: rapid hands-off workflows, flexible batch sizes, reduction of handoffs and simplicity of operation. He also noted the value in terms of organization that DNA extraction and library preparation can be consolidated onto a single, easy to operate instrument.


Volta is eagerly interested in what other protocols should be ported onto Callisto. Among protocols in development is Oxford Nanopore ligation libraries and bringing more steps of a hybrid capture workflow onto the instrument. I keep planting the idea of the Oxford Nanopore Ultra-Long prep, given the low shear forces, but was thinking poolside that an interesting add-on to the current HMW DNA prep would be to just run a standard ONT rapid reaction on the DNA after extraction.  What multi-step protocols do you think would benefit from some combination of little hands-on time, low shear forces, random access to the operating surface and miniaturization?  Perhaps with many steps and multiple bead cleanups? Perhaps you can dream a major Callisto workflow!

Volta isn't alone in trying to automate extraction and library preps, but there are far fewer solutions addressing the HMW/long read space than short reads. But others are out there. Integra continues to sell Miroculous, and I'm assured that this platform continues to be developed. PacBio just announced high throughput protocols for a number of conventional liquid handlers. As with all startups, Volta must demonstrate an ability to deliver reliable instruments, rapidly service any malfunctioning units and deliver reagents in a timely and consistent manner.

The long reed ecosystem has lagged the short read ecosystem by two thirds of a decade or perhaps more - to me. And it's clear to me, also to other early access customers including the Broad Institute, Mt Sinai in New York and the Hartwig Institute in Amsterdam, that VoltaLabs Callisto represents a significant step towards closing the gap in capabilities.

[2024-02-12 13:33 fixed a VoltaPrep to VoltaPure]


Anonymous said...

interesting post Keith.

are the electrowetting surfaces reusable or one-time use?
how much does each electrowetting surface cost?

Anonymous said...

If I remember correctly Otherwise than stated in the article the Miroculus MiroCanvas has never used oil!

Anonymous said...

Can this instrument preform protocols the require a PCR reaction?

Dale Yuzuki said...

Pretty neat tech!

IIRC the NeoPrep was only an 8-sample-at-a-time instrument, and the upshot was that you HAD to run 8 samples all at once. Also price-wise, the per-sample cost to do library prep was the same as if you ran it manually.

Also important to point out that Miroculus was only ONE sample at a time. Chatting with them at ASHG, it's about a 3 hour time for ONT or PACB library construction, but the stark thing that stood out to me (from a customer of theirs at Hopkins presenting a poster) was that you had to baby-sit the unit when it finished. You only had 15 minutes to grab the output sample, or else!

Really interesting technology - thanks for sharing!