Skip to main content
Premium Trial:

Request an Annual Quote

New Single-Cell Genome, Transcriptome Sequencing Technique Delivers Better Bang For Buck


This story has been updated from a version posted Mar. 14 to add context and detail to the results obtained in melanoma cells.

NEW YORK – A scalable and low-cost way to perform genome and transcriptome sequencing (G&T-seq) of single cells without upfront whole-genome amplification (WGA) is being developed at Belgium's KU Leuven.

The method, called Gtag&T-seq (Gtag), combines G&T-seq with tagmentation. Researchers from the lab of Thierry Voet at KU Leuven used it to improve coverage uniformity, lower selection and amplification bias, and show that transcriptome-based DNA copy number inference may not be as accurate as previously hoped.

In particular, the researchers used Gtag to describe for the first time, the heterogeneity of small focal amplicons between and within genomic amplicons and their potential role in tuning the melanoma cell’s response to treatment, highlighting the interplay with other genomic and epigenomic rewiring.

"We found a potential [cell state] switch," said Sebastiaan Vanuytven, a postdoctoral researcher at KU Leuven and one of the paper's co-first authors.

Their study is currently available as a preprint on BioRxiv and has been submitted to the journal Genome Research.

The Gtag technique consists of physically separating a cell's DNA and messenger RNA, then using tagmentation to produce a genomic DNA library, as opposed to preamplification via WGA. Tagmentation is a means of cleaving and tagging genomic DNA for downstream analysis without first amplifying it, thereby minimizing amplification and selection bias.

"Genomic tagmentation, although not completely random, is much more random than, for example, [Takara Bio's] PicoPlex" said Koen Theunis, a postdoctoral researcher at KU Leuven and the study's other co-first author.

Cell-specific barcodes and sequencing adapters are then added via PCR to enable multiplexed low-coverage single-cell sequencing.

"It's a nice study that really tries to solve a long-standing issue in single-cell sequencing — merging DNA and RNA information," said Billy Lau, a Stanford University researcher who works in single-cell profiling. "G&T-seq was really at the forefront of that, and it's nice to see updates to it."

In a direct comparison to G&T-seq using the HCC38 cancer cell line and its matched normal cell line, Gtag showed less noise in genomic readouts, and improved coverage uniformity and Gini index, a statistical method to identify genes whose expression varied least across a large set of samples.

Gtag also appeared to improve breakpoint and copy number calling relative to G&T-seq in a melanoma xenograft model.

The xenograft came from a melanoma patient with a mutation in the BRAF gene, and who had an almost complete response to therapy with combined therapy of a BRAF and a MEK inhibitor (dabrafenib and trametinib, respectively). The researchers collected cells from that xenograft before treatment with those inhibitors and at the point of minimal residual disease (MRD).

From these, they identified three major subclones, which showed 14 significantly differentially expressed genes between them and differed in treatment response and transcriptome plasticity over the course of therapy.

The subclones showed considerable heterogeneity in the presence, size, and amplicon copy number that the researchers say would have been difficult to resolve using bulk sequencing. Smaller amplicons and other small alterations, for instance, were identified via Gtag but not via PicoPlex, a technology used for comparison in profiling the xenograft subclones.

The different subclonal alterations led their respective cells to adopt distinct, drug-resistant phenotypic states. While some clones entered a terminally differentiated "pigmented" state, others adopted more invasive states.

"The combination of … genomic sequence and transcriptome sequence is really interesting because we can look into the functional consequences of mutations that we detect in the DNA of a cell," said Voet, a professor of medicine at KU Leuven and the study's principal investigator.

Interestingly, the group found virtually no transcriptomic differences between two proliferative subclones, despite one of them having lost a copy of chromosome seven.

"We were really amazed that even though you lose a whole copy of a chromosome, the effects were quite minimal," Vanuytven said.

To Stanford's Lau, the study confirms an observation made by numerous other researchers, that gene dosage, or copy number, and gene expression are not necessarily correlated with each other.

"You can't necessarily use transcriptome-based copy-number tools and get a reliable estimate of DNA states," he said, "and vice versa, you can't necessarily do the opposite and infer gene expression programs using purely DNA-based copy number. It's a really complex relationship, and I'm glad that there's another study pointing this fact out."

Todd Druley, Mission Bio's chief medical officer, also commended Gtag as a promising way to broadly examine both clonal and gene expression diversity.

"The authors have done a nice job improving upon whole-genome amplification of a single cell genome for accurate sequencing," he said. "This method has historically been fraught with error, but the authors show convincing benchmarking data."

Mission Bio is also active in the single-cell sequencing space and recently began work to establish proof of concept for a single-cell MRD test in acute myeloid leukemia.

Druley noted that beyond primary research, Gtag could find commercial applications, such as neonatal testing and MRD. He commented, however, that Gtag's potential clinical significance might be better ascertained through comparison to genome-wide arrays and/or in situ hybridization.

Druley also cautioned that "as with most preprinted manuscripts, one would expect there is additional key data waiting to be revealed in the final, peer-reviewed publication. As such, it's premature to suggest that this method, as presented, would be used for direct patient care."

Regardless of future applications, Gtag offers a highly economical alternative to other whole-genome sequencing techniques.

"For example," Theunis said, "if you sequence 384 cells with PicoPlex, the cost is about €8,000 (about $8,560), [wherein] the whole-genome amplification and library prep makes about 80 percent of the cost. For Gtag, the cost for sequencing 384 genomes is only €2,400."

Voet commented that while Gtag itself is not patented, there are already patents in place for several of the techniques that it relies on, such as single-cell haplotyping and single-cell DNA amplification and sequencing.

Voet declined to comment on any commercialization or licensing plans for the time being. He noted that the lab is investigating ways to perform Gtag at scale, such as by moving into droplet or nano well technologies. It is too early, however, to speculate on when to expect such expansions.

"It still requires a significant amount of work," he said, "to make this compatible with high-throughput analysis."