Skip to main content
Premium Trial:

Request an Annual Quote

DLP Plus Single-Cell Sequencing Method Spreads From Canada to Cancer Labs Worldwide


This story has been updated to include additional information about labs adopting DLP Plus.

NEW YORK – A recent method for high-throughput, single-cell, whole-genome sequencing is gaining practitioners worldwide, especially in the field of cancer evolution.

Initially developed at the BC Cancer Agency and the University of British Columbia, Direct Library Preparation (DLP) Plus is a platform designed to prepare single cells from cancer samples for low-coverage sequencing for copy number analysis, or CNA. It serves the same purpose as a predecessor method, simply called DLP, that used microfluidics to isolate cells. For the new version, the BC Cancer team ported that method to Takara Bio's nanowell plates, using cell spotting technology from French instrument maker Cellenion.

DLP Plus promises to have other applications besides single-cell genome sequencing. "We focus on a whole-genome sequencing approach because there's so much science to do there," said Sohrab Shah, a computational biologist who developed the bioinformatics side for the method at UBC before taking it to his new employer, Memorial Sloan Kettering Cancer Center. "But it should allow for full-length transcriptome sequencing, ATAC-Seq (assay for transposase-accessible chromatin by sequencing), and other types of assays," he said. "That's something that we should be able to accomplish in the future."

So far, researchers at BC Cancer have sequenced approximately 900,000 cells worth of data, according to Jecy Wang, a research assistant who works with the platform at the agency's single-cell sequencing core. "The goal is to do 1 million cells," she said, but it won't be long before they'll have to set new goals. Just a few years ago, they wanted to sequence100,000 cells. "It just jumps exponentially once you get the method going," she said.

The team is not trying to commercialize DLP Plus, but the platform has now been adopted at MSK, the New York Genome Center, the UK's Francis Crick Institute, and at Sweden's Lund University. As a genome-wide DNA sequencing method, it fills a niche without much direct competition. Its focus on the genome, rather than transcriptome sequencing, sets it apart from many droplet- or plate-based methods, such as 10x Genomics' Chromium or Drop-Seq. 10x used to offer a Chromium copy number assay, but discontinued it at the end of 2020. Other single-cell DNA sequencing methods are either targeted, such as Mission Bio's Tapestri platform, or use amplification, such as multiple annealing and looping-based amplification (MALBAC), or multiple displacement amplification-based methods, including BioSkryb's SkrybAmp.

DLP Plus, as it exists now, materialized from a chance encounter, according to Robin Coope, an instrumentation expert at the BC Cancer Agency Genome Sciences Centre who worked on the method and its predecessor. Hans Zahn, a former graduate student who was working on a grant proposal to scale up DLP, was in Coope's office and saw a Takara Smart Chip on the desk, a device that has more than 5,000 100-nanoliter wells.

"Can I take this with me?" he asked Coope. From there, he was able to implement the same tagmentation-based, PCR-free, Illumina library prep chemistry — which previously required microfluidics — in the open nanowells. "This is enormously easier to do than DLP," Coope said. "It makes it faster and more practical." Not only can DLP Plus do multiple assays, but it means the researchers are not limited to emulsion chemistries needed for microfluidics.

From there, optimizing the method took place in a lab called the "insanity room," according to Adrian Wan, a former lab manager who worked on the DLP Plus project. "There was just so much stuff to do" there, he said. The original DLP protocol required 10- to 12-hour days in the lab, which needed to be cut down to six to eight hours, and Zahn wanted to optimize the method for many different experimental conditions. The team also incorporated the Scienion S3 spotter, produced by a sibling company of Cellenion's, but ultimately replaced it with Cellenion's CellenOne spotter.

The Cellenion instrument can eject volumes as low as 0.5 nanoliter with high spatial accuracy, Coope said. The CellenOne can also look for cells in the nozzle, ensuring a higher proportion of single cells on the plate. "It's like the ability to see who is on deck, in baseball," Coope explained. "As you spot, the whole column of liquid moves down. It can image the end of the nozzle and tell if there's a single cell on deck, ready to go. If there isn't, or if there is more than one, it spots it into a waste bin."

Now, the UBC team can dispense about 1,000 cells in half an hour "if things are running well," Coope said. Cells from two or three samples are usually spotted on a chip for a total of about 3,000 per chip.

A thousand cells is also the sweet spot for the size of the pools of cells the researchers sequence, especially using the Illumina HiSeq X. That might change as they move to the NovaSeq platform, but it also helps keep the cost of sequencing down. "If you identify cells that are interesting, you want shallow pools that you can sequence," Coope said.

Still, the method can take up to four days, including cleanup and QC, especially with tissue samples. It works with fixed samples, too, though Wang said that is a new application they are still optimizing. Unsurprisingly, data quality is dependent on the quality of cells going into the spotter. "It has to be good," Wang said. "We're collectively getting better at understanding how to make cell preps that spot reliably."

DLP Plus comes with its own bioinformatics package. In addition to copy number analysis, it allows analysis of rearrangement breakpoints, helps build phylogenetic trees, and includes data visualization apps. Wan suggested that anyone interested in doing DLP Plus use the custom data package. "At the end of the day, sequencing data is sequencing data, but it might not make sense, because our pipelines are geared towards things we're looking at," he said.

"Typically, we can resolve events down to 500 kb at 0.1X coverage with 2 million reads," Shah said. "Obviously, there are diminishing returns as you ramp up sequencing depth, but we have sequenced some libraries up to 0.5X, or a little higher."

The cost per cell is approximately $.30 for reagents, or about $.70 including lab tech labor and the use of the machines, but Wan noted that this calculation was done in 2019.

So far, the main applications of DLP Plus are in cancer genomics. "We're a breast cancer lab," Wang said. "Everything was optimized based on breast tissues, breast tumors, and cell lines." But working with collaborators has also necessitated optimizing the method for other samples, such as colorectal tumors, fallopian tubes, and ovarian tumors. The lab has even run Arabidopsis samples from plant researchers.

The DLP Plus community, including Shah, Wang, and Sam Aparicio, who runs the single-cell core at BC Cancer, has already published a paper on clonal fitness based on data obtained with the platform. In June, Nature published their time-series analysis of polyclonal populations, based on 42,000 single-cell genomes.

They found that TP53 mutations "alter the fitness landscape, reproducibly distributing fitness over a larger number of clones associated with distinct CNAs," they wrote in the paper. In triple-negative breast cancer patient-derived xenografts, or PDX, models, they were able to "accurately forecast experimentally enforced clonal competition dynamics," and "drug treatment in three long-term serially passaged TNBC PDXs resulted in cisplatin-resistant clones emerging from low-fitness phylogenetic lineages in the untreated setting."

Shah said this approach can change the way that cancer evolution is studied by allowing the use of phylogenetics. "It can provide a fundamental backbone against which we can measure and model the evolution of individual clones," he said.

DLP Plus could also be applied to infectious disease evolutionary studies. "We can see how powerful it is in tracking disease variants in COVID-19," Shah said. "The same principles apply: If we're able to track clonal evolution in a patient, we could zero in on branches of phylogeny that are likely responsible for treatment resistance." Treatment could then be changed so the clone that has expanded is no longer favored.

"That's a very forward-looking view," he noted. "None of that is in place at the moment, but that’s where we’d like to go."

The publication describing DLP Plus came out in Cell in November 2019, just a few months before the COVID-19 pandemic began, so getting people trained on the platform has been difficult. The UBC team likes to bring people in because they know their in-house setup works. They've been holding training sessions for DLP and DLP Plus about once a year for the last three or four years, though they had to train New York Genome Center researchers via video calls due to COVID-19.

Despite the challenges, the team presses on. "Seeing that there are other places that want to adopt our methodology, it's very fulfilling," Wang said.