Skip to main content
Premium Trial:

Request an Annual Quote

At AGBT, 10X Genomics Presents Early Data, New Applications for Chromium System


NEW YORK (GenomeWeb) – 10X Genomics discussed details of its new Chromium system and presented some data from collaborations with a few early-access customers at last week's Advances in Genome Biology and Technology meeting in Orlando, Florida.

10X Genomics launched its first system, GemCode, at last year's AGBT meeting and in January, at the JP Morgan Healthcare conference in San Francisco, the firm announced that it planned to launch version two of its platform and an application for single-cell analysis. Last week, the company provided additional details on the single-cell technology and the second iteration of its platform, which it has dubbed Chromium.

The new system will cost $125,000, up from the $75,000 price of the original GemCode system, and will begin shipping late in the second quarter, CEO Serge Saxonov told GenomeWeb at the AGBT conference.

Customers will also be able to purchase whole-genome, exome, and single-cell kits, priced at $450 and $250 per sample for the genome and exome kits, respectively, and between around $.20 and $1.20 per cell for the single-cell kit, Saxonov said.

Chromium uses the same microfluidics and barcoding technology as the GemCode, but with greater resolution. The firm increased the number of partitions to 1 million from 100,000 and the number of barcodes to 4 million from 750,000. That means that instead of about 50 molecules of DNA per partition, there are now only five, which improves assembly and "helps get closer to single-molecule resolution," 10X CSO Ben Hindson told GenomeWeb.

The new system will also be compatible with Illumina's HiSeq X Ten so customers can use it for large-scale population studies.

With the GemCode, the firm recommended that users create two libraries — a standard Illumina library and also a 10X-specific library — but with the Chromium, only one 10X library is needed.

During a presentation at the AGBT meeting, Stacey Gabriel, director of the Genomics Platform at the Broad Institute, discussed her institution's experience with 10X's technology. The Broad received the GemCode platform in June 2015 and has been an early-access user of the Chromium since December, she said.

The institute has been testing whether a single library system is adequate to call both SNVs and structural variants using Chromium, and evaluating the instrument's workflow, potential biases, and its ability to phase genomes.

So far Broad researchers have tested Chromium on 30 samples, including reference trios from the Genome in a Bottle Consortium, cancer cell lines, and autism cell lines.

The workflow has been straightforward and the Chromium library prep is an improvement over the GemCode library prep because the DNA shearing step has been removed, she said.

Whole-genome sequence data generated from the 10X platform had similar mean coverage, coverage distribution across the bases, and expected insert sizes to Illumina sequencing with the Broad's standard whole-genome protocol.

One notable improvement of Chromium compared to GemCode was GC bias, she said. While the GemCode libraries displayed a fair amount of bias, the Chromium libraries were "pretty smooth — nearly as smooth our PCR-free libraries," she said.

Variant calling was also "good out of the box." she said.

The Broad team also tested the ability of the exome kit to phase samples. Daniel MacArthur, associate director of medical and population genetics at the Broad Institute, sent some "difficult" samples to 10X Genomics for exome library preparation using Chromium, Gabriel said. The goal was to see whether the technology could phase compound heterozygous mutations in a cell line. 10X Genomics researchers used Chromium and Agilent's SureSelect exome kit to prepare exome libraries from a cell line that had two causal mutations in the Titin (TTN) gene, a notoriously difficult gene to sequence due to its size and repetitive regions.

Gabriel said that the data it received from 10X Genomics showed that it had phased the two pathogenic variants in TTN that were 189 kb apart. 10X also phased the entire exome into haplotype blocks with an N50 of 245 Mb.

For other samples, 10X phased variants that ranged from 30 kb apart to over 50 kb apart using its exome kit, Gabriel said.

10X Genomics and Agilent are also codeveloping a "premium" exome, the companies said last week.

Gabriel said that Broad researchers Marcin Imielinski, Cheng-Zhong Zhang, and Matthew Meyerson are working with 10X to evaluate complex cancer genomes. She described one case in which the group was able to use the linked reads to identify break points in copy number alterations and another case where they were able to use the linked reads to reconstruct the path of a complicated series of rearrangements.

Chromium is a "significant improvement" over GemCode, Gabriel concluded, and said that the Broad team planned to use it in its studies on rare disease, cancer, and common diseases. In addition, she said that the team would continue to work on improving assembly algorithms for Chromium and to work on increasingly large projects, eventually population-scale studies.

The HudsonAlpha Institute of Biotechnology will also be an early-access user of the Chromium system. Shawn Levy, director of HudsonAlpha's Genomic Services Lab, said that the institute has been running the GemCode platform since around June and has received libraries generated by 10X Genomics on the Chromium that it has subsequently sequenced at its institution on Illumina's HiSeq and HiSeq X Ten.

On average, he said, the Chromium libraries have generated between 860 million and 884 million reads per lane and the genomes were sequenced to a mean depth of 32x to 34x with an insert size of 350 base pairs. PCR duplicates were between 5 percent and 6 percent and 99 percent of the SNPs called were phased in haplotype blocks with N50s averaging between 7 Mb and 9 Mb. The longest haplotype block was 30 Mb.

Levy said that the linked reads have made it easier to resolve structural variants. For instance, in one of the first genomes it sequenced, there was a copy number change that "wasn't completely obvious from just the HiSeq" whole-genome sequencing data, but readily detectable with the linked reads from 10X Genomics. In another case, he said, the linked read data helped to resolve a tandem duplication that was "completely invisible" with standard sequencing data.

Although the 10X Genomics data does not enable de novo assembly of species without a reference, Levy said that the HudsonAlpha team demonstrated that it could de novo sequence a clouded leopard genome and map it to the cat reference to identify interesting structural information.

Levy added that the institute has not yet evaluated other aspects of Chromium like potential biases or artifacts in the data. Primarily it has been using it to push the limit on structural variant detection, he said.

Single-cell RNA-seq

One application that 10X plans to make available even before the full commercial launch of the Chromium is single-cell RNA sequencing. 10X Genomics will begin shipping the single-cell kit in March, and it can run on the GemCode platform. For customers that plan to order a Chromium system, the single-cell capability can be swapped when the Chromium system becomes available, Hindson said.

Initially, customers will be able to load between 1,000 and 6,000 single-cells per channel and can run all eight channels in parallel, for up to 48,000 cells, Saxonov said.

As the firm previously described, single cells are loaded onto the platform instead of DNA. RNA is then extracted and molecules from the same cell are barcoded. After sequencing, reads are grouped back to their respective cells.

In a collaboration with Jason Bielas from the Fred Hutchinson Cancer Research Center, researchers tested the single-cell RNA-seq capability on cell lines. Bielas told GenomeWeb/said during an AGBT presentation that validation experiments demonstrated that the technology could identify subpopulations of cells even when they constituted only 1 percent of the population.

From data generated by 10X scientists, Bielas said that single-cell RNA-seq found that the expected subtypes of a population of peripheral blood mononuclear cells — such as naïve T cells, CD4 T cells, dendritic cells, NK cells, and progenitor cells — were all recoveredat the expected proportions.

Bielas added that he plans to use the single-cell RNA-seq capability to study more complex oncology samples such as chronic lymphocytic leukemia samples to identify markers of relapse, and also in leukemia patients who have received bone marrow transplants to study immune reconstitution at the single-cell level.