Skip to main content
Premium Trial:

Request an Annual Quote

Genapsys Publishes Sequencing Data but Questions Linger About Technology


SAN FRANCISCO (GenomeWeb) – Genapsys, which aims to launch a portable sequencing platform that uses electronic detection in the second half of this year, has published data from a number of different samples in a preprint in BioRxiv, benchmarking the system's performance.

In the study, the company described the technology's performance using a human cancer cell line, a human cell line reference sample, and microbial genomes, performing targeted sequencing, exome sequencing, and whole-genome sequencing on the samples, respectively.

The company performed a "good battery of tests to demonstrate the utility of the platform," said Chris Mason, an associate professor at Weill Cornell Medicine who was not involved in the study. "Most of what we do when we benchmark a technology, they did those tests. So it was a good demonstration of the platform on multiple modalities and applications."

The researchers first described sequencing an Escherichia coli genome. The average read length was 141 bases and average coverage across the 4.69 Mb genome was 320X. In total, they generated more than 12 million reads, 89 percent of which could be mapped. The average cumulative accuracy at base position 100 was 99.8 percent. More than 97 percent of the bases had a Q score greater than 20, while around 64 percent had Q scores greater than 30.

The researchers also analyzed microbial genomes with different GC content to characterize the amount of GC bias in the platform, including Campylobacter jejuni, with an average GC content of 30 percent, and Bifidobacterium animalis, which has an average GC content of 61 percent.

The preprint also described the results of experiments the company had previously discussed at the Advances of Genome Biology and Technology conference earlier this year, including the sequencing of the well-characterized NA12878 human cell line, which has served as a human reference standard for the Genome in a Bottle Consortium.

For that sample, the team sequenced the exome to 96X coverage using Integrated DNA Technologies' probe-based capture kit. They identified more than 20,000 SNPs, which they compared to Illumina sequence data, which identified just over 21,000 SNPs.  There was a high level of overlap between the two sets of SNPs — 132 SNPs were found only in the Genapsys data and 735 SNPs only in the Illumina data. When compared to the GIAB high-confidence call set, the Genapsys system had a sensitivity of 95.6 percent and a precision of 99.0 percent.

The researchers also analyzed a cancer sample using a targeted pan-cancer hybridization capture panel from IDT. They analyzed a sample that contained a mixture of DNA from three cell lines with known mutations and allele frequencies between 1 percent and 24.5 percent. Sequencing generated around 11.5 million reads, about 64 percent of which were on target. Average coverage of the targeted regions was 624X.  The researchers noted that they were able to correctly identify even the low-frequency mutant alleles and that the detected variants and variant frequencies correlated well with both Illumina sequencing and droplet digital PCR.

"Overall, it's intriguing, but there are some things to keep an eye on," Mason said. Cost per gigabase, throughput, ease of library prep, multiplexing capacity, and scalability are all lingering questions, he added.

Mason noted that previous sequencing technologies that are based on CMOS sensors have had trouble scaling. For instance, he said, Thermo Fisher Scientific's Ion Torrent had difficulty when it tried to scale its Ion Proton system, which also uses a CMOS sensor-based chip. Increasing, the density of the chip turned out to be more problematic than the company originally anticipated, he said, and ultimately, the firm never commercialized a long-awaited second version of the chip. Nonetheless, the performance demonstrated in the preprint was good, particularly for a first version of the instrument, he said.

Genapsys has not yet disclosed details of its technology, such as its library prep process or details of its CMOS detection method. For library prep, CEO Hesaam Esfandyarpour said only that is uses clonal amplification that is not based on PCR and the preprint itself does not contain additional details.

The technology detects base incorporation using an electronic signal. The amplified DNA molecules sit within a detection chamber along with CMOS sensors. Nucleotides are injected into the space one at a time and if the nucleotide binds to the DNA molecule, the sensor detects the incorporation.

"One thing that's frustrating about the preprint is, there is a lack of technical detail," said Keith Robison, a principal scientist at Ginkgo Bioworks who also analyzed the Genapsys preprint on his 'Omics 'Omics blog. "It's hard to know what they're doing, which is frustrating for someone trying to follow the field," he said. The library prep process in particular would be important to understand, since that will have an operational impact for the end user, he added.

Robison said that the instrument cost, anticipated to be around $10,000, and the footprint of the system made it very appealing, but if library prep ended up being too onerous or involving a lot of ancillary equipment, that would negate much of those advantages.

Mason said he thought the technology seemed most similar to Illumina's iSeq instrument, which also uses a sequencing-by-synthesis chemistry coupled with CMOS-based detection. The iSeq, however, is able to do paired-end sequencing with 250-base reads. Mason said that cost per gigabase for sequencing on the iSeq was around $300, depending on the specific sequencing run, so it would be important to compare running costs of the systems, as well.

Genapsys has said it is developing three versions of its chip: one with 1 million sensors, which will cost $100 per run; a second with 16 million sensors, which will cost $300 per run; and a third with 144 million sensors that will cost $600 per run. In the preprint, the researchers reported that one run on the instrument, sequencing a microbial genome, generated around 1.5 gigabases of data. The company generated that data using its medium-throughput $300-per-run chip, resulting in an estimated cost per gigabase of $200.

Mason also noted that the applications for both the Genapsys instrument and the iSeq would likely be similar: rapid QC, panel testing, basic metagenomics, and potentially infectious disease diagnostics.

"I'm curious to see where [Genapsys] takes it," he added. "It's always good when there's more competition because it pushes the market."

The instrument is not yet commercially available, but Esfandyarpour said the company is on track to launch it in the second half of the year. In addition, it has placed instruments with at least three early-access users, at the HudsonAlpha Institute for Biotechnology, Stanford University, and the Jackson Laboratory. The Jackson Lab's Chia-Lin Wei previously said that her group was testing a 212-gene cancer panel on the instrument using the mid-throughput chip, but had not at the time done a thorough analysis of the system. The other early access users have not responded to a request for comment on their experience with the Genapsys platform so far.