NEW YORK (GenomeWeb) – Scientists interested but not well versed in transcriptome profiling can now go to the Broad Institute for help. The Broad Technology Labs Specialized Services Facility offers automated services for high-throughput characterization of both single cells and populations of cells.
Both services are based on high-throughput 3' digital gene expression protocols Tarjei Mikkelsen, a scientist at the Broad Technology Labs, told GenomeWeb. "It's based on an old concept that you can measure gene expression by sequencing tags or small fragments from each transcript in a cell to measure expression levels of genes," he said, noting the contrast to RNA-seq methods where the goal is to sequence an entire transcriptome or to get splicing info.
Single-cell RNA barcoding and sequencing (SCRB-SEQ) and high-throughput digital gene expression (HT-DGE) are essentially the same method, Mikkelsen said. "We first developed SCRB-SEQ but we realized pretty quickly it's the same steps, and the same method was applicable to more than one cell."
The protocol is based on automated handling of samples using standard 384-well plates. In SCRB-SEQ, the cells are sorted into each well; in HT-DGE, cells might be cultured first or RNA might be extracted from tissue samples before being deposited on the plates.
Cells are then lysed and a reverse-transcription step to create cDNA incorporates a unique barcode for each well. Once the cells are barcoded, they're pooled into a single tube. The lab processes the cDNA to capture the 3' ends of transcripts, and then those are sequenced in order to count the number of different transcripts for each gene and each sample, Mikkelsen said.
"The primary reason for using this protocol is to simply get a count of how many transcripts there are for a given gene in each sample," he said. There's some additional information available, but he said the method won't provide a characterization of splice variants or mutations within transcripts. "It's about getting gene expression values," he said.
Tagged sequencing is not a new technology, Mikkelsen acknowledged. "There are similar flavors out there, but we've optimized our protocol to minimize sample cost and maximize throughput and make it easy to automate the protocol we're running. We've made very pragmatic and practical improvements."
The tag sequencing is optimized for counting, he said, and the lab came up with a simple scheme for extracting the 3' ends of transcripts using hybrid PCR to minimize the number of steps needed relative to earlier protocols, which increased throughput. Another major conceptual advancement is that SCRB-SEQ and HT-DGE use so-called early pooling methods. "As soon as cDNA has been tagged, everything is pooled together, which minimizes the number of reactions we're handling. That has benefits of being efficient if you have a large number of samples," he said.
Late pooling is a different option that can be used to make complete libraries from each sample and pool at the very end, Mikkelsen said. "If you're handling heterogeneous samples, there's a bit more room for normalizing the libraries depending on yield. Early pooling methods are more dependent on having the input samples being normalized."
The services are offered in addition to sequencing to both internal and external customers, both private and academic. Mikkelsen declined to say how much they cost, but noted that both are dependent on batch and sample sizes. "The more plates we can process, the cheaper the unit cost will be," he said. The turnaround time depends on demand, but he the protocol itself can yield sequencable DNA "within a day or two."
While the lab has offered these services for almost two years, Mikkelsen said they're always looking to increase throughput. They've worked with commercial vendors to explore ways to use the same molecular biology on different plate formats and technology.
Collaborating with WaferGen Biosystems on single-cell RNA sequencing has led them to implement the same RNA-seq protocol on the nanoliter chip. Mikkelsen said they've had similar collaborations with Fluidigm on that firm's C1 system, among other collaborators /business-news/wafergen-signs-two-additional-early-access-partners-single-cell-analysis-tech
"There are many different ways to achieve the same result, lots of different small variations in protocols that exist out there," Mikkelsen said. "We're getting to a point where variations are understood well. RNA sequencing as a concept isn't a new thing. Success or failure of the methods will be whether good science comes out of them."
Mikkelsen said that he's seen demand for the services grow. "We see a lot of different use cases: people collecting clinical samples and exploratory parts of clinical trials, people collecting large tissue samples and biopsies. There has been a lot of interest from chemical screening labs. Screens might have several dozen 384-well plates exposed to different compounds and low-cost RNA-seq is competitive with other methods to phenotype those cells," he said.
New kinds of labs are becoming interested in gene expression profiling and Mikkelsen said it's often more cost efficient for them to outsource library construction to someplace like the Broad Technology Labs rather than train their own staff and invent their own protocols.
"There's been a lot of interest from labs that are really focused on cell biology or high-throughput functional assays and have been holding off on going into [RNA-sequencing]."
Mikkelsen said there seems to be a sweet spot for labs looking for rich phenotyping of samples in a range between hundreds of samples and and thousands of samples. "RNA-seq is becoming a good alternative for getting that at a simple cost," he said.