NEW YORK – Scale Biosciences, a startup pursuing commercialization of technologies that increase the scale and throughput of single-cell and spatial biology experiments, has begun inching out into public view. The firm, which has not officially launched yet, plans to sell kits employing single-cell combinatorial indexing (sci) technology developed in the lab of University of Washington professor Jay Shendure, GenomeWeb has learned.
Earlier this month, three ScaleBio employees published a paper in Nature Biotechnology in collaboration with researchers at Oregon Health & Science University led by Andrew Adey — a former postdoc in Shendure's lab. The article describes a method to increase the number of reads per cell obtained in sci-ATAC-seq (assay for transposase-accessible chromatin by sequencing) 200 times.
One of the authors is Frank Steemers, the former director of Illumina's advanced research group, who is VP, CSO, and a cofounder of ScaleBio. He and Shendure are not the only attention-grabbing names on the firm's list of founders, which also includes Stanford University professor Garry Nolan, who also cofounded Akoya Biosciences; CEO Sean Scott, Qiagen's former chief business officer; and UW professor Cole Trapnell, who has collaborated with Shendure to develop combinatorial indexing-based assays. Shendure was also a cofounder of Bellwether Bio, a liquid biopsy company that Guardant Health acquired in 2019.
The paper, originally posted as a BioRxiv preprint in January, is just the first peek from under the veil for the firm. The company has also joined the public membership rolls of the California Life Sciences Association (CLSA), a biotech industry group, and has posted about half a dozen job ads.
Records kept by the state of Delaware show that a company named Scale Biosciences was incorporated there in December 2019. According to Pitchbook, a firm that collects and information on private capital markets, Scale Biosciences has raised $32 million in Series A financing, with participation from Tao Capital Partners. Tao did not respond to a request for comment.
According to the spring 2021 newsletter from CLSA, ScaleBio joined as a gold-level member sometime between August 2020 and February 2021. The address on file with CLSA places the firm in San Diego, specifically at Johnson & Johnson's JLabs startup incubator campus. JLabs said in an email that it was "unable to provide" information about the company. The Nature Biotechnology paper listed Berkeley, California as the ScaleBio team's location.
In addition to the paper's coauthors, which include Dmitry Pokholok, a former Illumina scientist, the company has at least two other employees, a bioinformatics lead and a market development lead, according to their LinkedIn profiles.
In its own words, ScaleBio is a "life sciences tools company offering 'cell-to-insight' single-cell sequencing library preparation solutions compatible with and independent of on-market single-cell systems for the research and clinical markets," according to a description on the BioSpace jobs board, which notes that its "patented combinatorial indexing technology and methods enable increased sample indexing, cell throughput, and multiomic profiling of cells at scale and at low cost."
In single-cell analysis, 10x Genomics is the current market leader, though that firm is competing with Bio-Rad Laboratories, Parse Biosciences, Singleron Biotechnologies, and others. Parse, a UW spinout formerly known as Split Biosciences, also uses a combinatorial indexing scheme but the way barcodes are attached to transcripts is different from ScaleBio's, according to Parse.
In spatial biology, a slew of companies are trying to develop technologies that offer data on expression of any number of genes, up to the entire transcriptome, at varying levels of spatial resolution, even identifying transcripts a the subcellular level.
Since 2014, Shendure, Trapnell, and Adey, often in collaboration with Steemers, have been developing combinatorial indexing, the basic premise of which is to attach successive rounds of oligo barcodes onto pieces of DNA at various steps, enough to statistically guarantee that each molecule in a sequencing library has the cell or sample that it came from encoded by a unique combination of barcodes. One round of barcoding usually happens during tagmentation, though they can also be added during PCR amplification.
The group has developed "sci" flavors of many assays popular in the single-cell field, including sample multiplexing, RNA sequencing, Hi-C chromatin conformation assays, ATAC-seq, and spatial methods that analyze single cells (sci-Space) or tissue microbiopsies (sci-Map.) Some assays combine two approaches at once. Exactly which of these will be commercialized by ScaleBio is unclear. The company, Shendure, and Trapnell did not respond to a request for comment before deadline.
The technology described in the Nature Biotechnology paper could be the key to unlocking the full potential of these combinatorial indexing assays, which all employ some sort of tagmentation with Tn5 transposase. More or less all previous tagmentation-based assays randomly incorporated forward and reverse sequencing adapters into pieces of DNA. Because half the library molecules would have two adapters of the same type, "already, the theoretical maximum efficiency is only 50 percent," Adey said. "This new workflow is a way of getting around that cap."
The new approach takes advantage of some of the transposase's quirks to do tagmentation with only forward adapters and then adds reverse adapters with a different biochemical process, so every molecule is viable for sequencing.
In the paper, the OHSU and ScaleBio researchers used the chemistry to improve the number of reads per cell in sci-ATAC-seq, sci-Hi-C, and sci-genome sequencing. "But it can be applied to any assay that uses tagmentation," Adey said, even bulk ones.
For their new assay, called symmetrical strand sci-ATAC-seq (s3-ATAC-seq), the researchers compared their assay to data generated with single-nucleus ATAC-seq, a method developed by University of California, San Diego researcher Bing Ren, and to 10x Genomics' single-cell ATAC-seq assay. "We get at least six times higher coverage than any dataset that has ever been publicly released," Adey said.
Theoretically, one could expect a twofold improvement by fixing the adapter problem, but because the Tn5 enzyme binds DNA so tightly and inhibits PCR amplification, the original assays are not fully efficient.. "This adapter switching step also sets up the PCR reaction to be super efficient," Adey said, resulting in the actual yield to be even better, compared to previous assays.
For single-cell DNA sequencing, Adey said his team could get a 10-fold increase in reads per cell just from their new adapter switching strategy, and up to 200-fold improvement with other optimizations. That level of coverage is comparable to what one could expect from flow sorting a cell into a well, except the "sci-" methods are much higher throughput. For example, in sci-seq proof of concept studies, the researchers profiled more than 15,000 single cells. Adey added that, theoretically, this chemistry could also be applied in droplets. "We haven't explored that yet because plate-based methods are super cheap to do," he said.
The boost in reads per cell has important consequences for experimental design. "For one, you can get higher resolution of cell types from fewer cells," Adey said. "We're applying this to samples like tumor biopsies, where we can hardly get any cells out of any single-cell assay and we need to get as much as we can from each cell.
For single-cell ATAC-seq, specifically, this approach provides access to many more accessible sites. "In a typical assay, you're profiling maybe 10 percent of accessible sites," Adey said. "If you really want to look at what's active in that cell at the same time, there's a lot of missing information … We can actually start to ask questions about regulatory element interactions within individual cells."
Adey said that this chemistry was not the only improvement to tagmentation-based assays his lab has devised. "We've got some other ones that work better for other assays that I can't go into detail on," he added.
Those approaches could very well find their way into the assay kits that ScaleBio is seeking to develop. The Nature Biotechnology paper disclosed that Adey and Steemers are authors on a patent covering the technologies described therein. Adey said that ScaleBio has licensed those patents from OHSU but noted that he does not have an ownership stake in the company and, aside from research collaborations, will not be involved with it.
The firm is hiring staff to develop kits as well as manage beta- and early-access programs. Job ads state that its technology will work with existing droplet- and plate-based single-cell technologies, and that it will also enter the red-hot spatial analysis market.
Shendure, Trapnell, and Steemers have been listed as co-inventors on US and world patent applications, including one titled "High-throughput single-cell libraries and methods of making and of using," which has listed the University of Washington and Illumina as assignees. Whether ScaleBio has a relationship with Illumina is also unclear. Illumina and UW CoMotion, the school's technology transfer arm, did not respond to a request for comment.