By Monica Heger
Researchers at the Riken Omics Science Center in Japan reported recently on a sequencing technique to analyze gene expression on the Heliscope, and are currently using the method on the center's four HeliScope machines in the Functional Annotation of the Mammalian Genome, or FANTOM5, project.
The goal of FANTOM5 is to map the transcription start sites of every human cell type in order to piece together transcriptional networks.
The FANTOM consortium is an international research project that was initiated by Riken in 2000. The first four phases of the project focused on identifying the transcribed components of mammalian cells, and researchers cloned and annotated more than 100,000 full-length cDNAs from mouse.
The fifth phase of the FANTOM project, which is focused on human cells, kicked off last December. The main technology used for the project is an in-house developed method called HeliscopeCAGE, which the researchers are using to study transcriptional start sites in primary cells.
The technique, described in Genome Research last week, is based on a previously published method known as cap analysis of gene expression, or CAGE, except that it has been modified for a Helicos sequencer.
Riken researchers originally published CAGE in 2003, and last year developed a version called nanoCAGE that requires only 10 nanograms of starting RNA (IS 6/15/2010).
The initial versions of the method were originally run on the Illumina sequencing platform, but the iteration developed for the single-molecule Helicos system has been simplified to avoid second-strand synthesis, ligation, digestion, and PCR. Additionally, while the original version uses 50 micrograms of total RNA, the recent version requires only 5 micrograms, and a lower-quality version can work with just 100 nanograms of RNA.
"The main difference [from CAGE], is that the method is much simpler; we just sequence first strand cDNA," said Alistair Forrest, a co-author of the paper and leader of the Genome Profiling Technology Unit at the Riken Omics Science Center, via e-email.
While this is significantly more starting material than what is required for nanoCAGE, the new technique does not require any amplification steps, reducing bias, explained Piero Carninci, who has helped develop all iterations of the CAGE protocol at Riken and is involved in the FANTOM5 project.
"The more PCR cycles we use, more biases are introduced, which require more replicas in order to draw accurate conclusions on expression," said Carninci via e-mail.
He noted that the researchers decided to use the Heliscope for FANTOM5 because it provides "the least possible bias to quantify CAGE tags." This is an important feature, he said, "particularly to reconstruct transcriptional networks based on transcription starting-site centered expression profiling."
In the Genome Research study, the team demonstrated the method on a human myeloid leukemia cell line, generating three libraries, each of which produced 12 million to 18 million mappable reads.
This is superior to the previous version of the method that had been developed on the 454 GS FLX and only produced 1 million useable tags per sample with shorter average tag lengths.
Additionally, the 454 protocol "required 50 micrograms of total RNA, took multiple weeks to prepare, multiple rounds of PCR amplification and concatenation," the authors wrote, compared to the Heliscope method, which required no amplification and took one to two days per sample to prepare.
To assess even rarer cell populations, they tested starting with as little as 100 nanograms of RNA, which Carninci said is the "current limit of the technology." The lower input method detected expression for about 60 percent of the 13,468 loci detected with the 5-microgram library.
The method is the primary technology being employed for the FANTOM5 project, said Forrest. The project involves "profiling as many different biological states as possible," and will "give us a much better understanding of how promoters work," he added.
Eventually, Carninci added, the project's goal of mapping transcriptional start sites in primary cells will help piece together transcriptional networks, which could lead to an understanding of how cells are made and could be used for cell reprogramming. Currently, no other sequencing project is looking at the transcriptional start sites of primary cells, he added.
The FANTOM project still "has the character of basic science," he said, but, "ultimately, it will help to understand how [primary] cells are made."
He said there is not yet a set time frame for the project, but said the group hopes to publish several papers within the next couple of years.
Forrest added that the team has already used HeliscopeCAGE to analyze thousands of human and mouse primary tissue samples to "map mammalian promoters and collect an extensive promoter level expression atlas."
That data will be used to build "transcriptional regulatory network models for the majority of mammalian cellular states," he said.
Despite stopping production of its sequencing machines, Helicos is still providing reagents for the Riken team's four HeliScopes, which are all located in Japan. Additionally, the company provides sequencing services when extra capacity is needed, Carninci said.
In addition to the HeliscopeCAGE method, the team is using Illumina sequencers for RNA sequencing, ChIP sequencing, and other methods to generate supporting data for the project, Carninci said.
Have topics you'd like to see covered by In Sequence? Contact the editor at mheger [at] genomeweb [.] com.