Countable Labs, formerly Enumerix, was founded by serial entrepreneur Stephen Fodor, who originally stormed on the molecular tools scene with Affymetrix. I caught up with their new CEO, Giovanna Prout, at ASHG the other week and got a rundown on their new approach to counting molecules with PCR.
Prout recently found herself out of the CEO job at Scale Biosciences after its acquisition by 10X Genomics. She says she resolved to become the best stay-at-home mother ever and her kids loved having her home - but soon urged her to find a new gig as they recognized it was what made her happiest. So she quickly landed the CEO role at Countable Labs. Formerly Enumerix, the company is yet another molecular tools company from prolific scientific entrepreneur Stephen Fodor, best known for Affymetrix.
Countable’s standard workflow is simple. A 50 microliter reaction of sample DNA, probes, primers, PCR mastermix, and Countable’s proprietary matrix consumable are placed in a spin column. Centrifuging the columns generates a matrix, with approximately 30M individual picoliter-scale compartments capturing individual DNA molecules. After a brief (60 minute) PCR amplification in a conventional thermocycler (Countable has a list of preferred instruments), the tubes are placed in Countable’s benchtop instrument for light sheet microscopy imaging of the compartments, requiring 5 minutes of imaging per tube. There’s no dead volume - every picoliter scale chamber in the tube will be imaged. Because there are so many compartments, the system has a dynamic range of 6 logs! Countable’s instrument holds 96 tubes in the form of 24 strips of 4 tubes each. The instrument is priced at $150K with consumables adding up to $16 per sample. A full set of 96 samples can be processed in half a workday.
Assays can be designed using Countable’s universal multiplexing kit or conventional TaqMan hydrolysis probes can be included. The universal multiplexing kit has the advantage of enabling the use of inexpensive, fast arriving ordinary oligos which simply require a 5’ tail sequence on the forward primer to enable linking (via primer extension) the universal multiplexing codes to the user primers. Countable provides a software tool to streamline converting existing assays into universal multiplexing assays and analyze the resulting multiplex primer designs for undesirable cross-reactivity. Assays based on the universal multiplexing kit can be designed and tested in under a week
Countable is developing a high degree of multiplexing by managing the optical system so that it is capable of distinguishing 10 different fluorescent dyes. By imaging in 9 different channels, each channel a different pairing of excitation wavelength and emission wavelength, each dye can be distinguished by the unique fingerprint of intensities in each channel. This yields 10 clearly separable labeling schemes. Theoretically 48 different labeling schemes can be distinguished in this way. One ASHG poster from Countable demonstrated 8-fold multiplexing
So what can you do with so many colors? And with 6 logs of dynamic range?
One poster presented by ASHG focused on BRAF oncogenic mutation detection. Three clinically relevant mutations are seen at position 600 of the protein: V600E, V600K, and V600R. By first using 18 cycles of a PCR design agnostic to the status of codon 600 as a pre-amplification and then using allele-specific primers covering the four alleles (three oncogenic variants plus wildtype) and each primer a different color, Countable was able to demonstrate detection of the variants when present in a sample at a frequency of 0.08% - much better than the 1-5% achievable with qPCR and 0.1% for digital PCR. That’s also not a trivial detection limit to achieve with a sequencing assay - but at about $16 per sample the Countable assay will be far less expensive than any NGS assay unless you can batch to a very high degree. These results also leverage the high dynamic range of Countable’s assay - detecting between 400 and 700K molecules with a single assay system.
In another poster, Countable demonstrates measuring mitochondrial genome copy number, using multiple distinguishable probes targeting the mitochondria plus an additional one to get the nuclear genome as a reference.
Countable is also touting that they have built the system for GMP workflows, with the built-in software providing audit trails and other required security features for 21 CFR Part II compliance.
Another interesting feature of Countable PCR is the ability to recover samples post-amplification - a protocol is provided to extract DNA out of the matrix. So you can count from a precious sample and then potentially fully sequence it as well.



2 comments:
Nice tech, thanks for writing up Keith. Any idea how do they get rid of excess primers? Also, why is the sensitivity at 0.08%? If they have 6 logs then much better sensitivity should be possible isn't it?
Hi there- thank you for your great question. I'm a representative from Countable Labs.
With regards to the primers: Excess primer removal is unnecessary in Countable PCR. The reaction format ensures that each compartment interrogates a single target, and only amplified products contribute to the digital signal. Unamplified primers do not generate signal, eliminating the non-specific amplification artifacts that you usually see in other methods.
With regards to the sensitivity: The 0.08% sensitivity reported in the KRAS poster reflects several practical considerations:
1. Input material limitations: In cfDNA applications, you're typically working with only a few thousand DNA molecules as input. This creates a practical sensitivity ceiling around 1 in 3,000 molecules (~0.03%), regardless of the assay's theoretical dynamic range.
2. Optimization status: The KRAS assay shown in that poster had not yet been optimized with pre-amplification. With pre-amp optimization (now implemented), we've improved sensitivity significantly.
3. Context of the data: The poster data used synthetic cfDNA spike-ins with ~140,000 wild-type molecules as denominator. With pre-amplification and optimized conditions, we can now achieve sensitivity approaching 0.02% (200 mutant molecules in 900,000 wild-type background) while staying above pre-amp noise floor.
4. Sample type matters: The sensitivity achievable varies by application and input abundance:
--cfDNA: Limited by input (~few thousand molecules) → practical limit ~0.03%
--gDNA: With more abundant input, we've achieved 0.003-0.004% VAF LOD depending on the assay
If you have any further questions, please feel free to contact us: https://countablelabs.com/contact
Post a Comment