Systems biology boosted by RNA-sequencing consortium.
Heidi Ledford
A range of transcription factors are involved in setting a maturing white blood cell on the right path.MEDICALRF.COM/CORBISAn international consortium has released an analysis of unprecedented detail showing the genes and proteins that guide an immature cell to its final identity. The compendium of data could boost efforts to model the molecular networks that determine cell type — a long-standing goal in systems biology, and possibly a crucial step to creating stem-cell therapies.
The FANTOM consortium used high-throughput sequencing to create a timeline of the messenger RNA molecules produced by human leukaemia cells in response to the presence of phorbol myristate acetate. The chemical prompts the cells to stop proliferating and instead differentiate into mature white blood cells. The team took samples at six time points before and during the differentiation, and sequenced enough RNA molecules to allow detection of very scarce material found in as few as 1 in every 100 cells.
The first analysis of this RNA data, published online this week in Nature Genetics1, uses them to reconstruct the network of proteins that turns genes on or off as a cell takes on its final identity. Although similar models have been constructed for yeast, the FANTOM project is the most in-depth study to date of human gene regulatory networks, says Eran Segal, a computational biologist at the Weizmann Institute of Science in Rehovot, Israel.
The techniques developed by FANTOM should also provide a valuable resource for other systems-biology endeavours, such as modelling responses to drugs or learning how to coax a stem cell to become a neuron.
The FANTOM team, a collaboration involving more than 100 labs, took the RNA sequence data and compared them to full genome sequences to identify the sites of the genes being transcribed. The group then searched regions of DNA upstream from those genes for places that looked to be likely binding sites for proteins called transcription factors, thus showing which of those transcription factors might be playing a part. The involvement of 52 transcription factors was confirmed experimentally by using RNA interference, which can reduce the level of a given protein in a cell. When the production of any one of those 52 transcription factors was reduced, the differentiation process was perturbed.
The models show that a complex network of transcription factors is responsible for a cell's differentiation, with no one 'master regulator' in control. "It's like a transcription-factor democracy," says Harmen Bussemaker, a computational biologist at Columbia University in New York. From an evolutionary standpoint, distributing responsibility is a good strategy, he says: "It would not be a good design principle to have an Achilles' heel."
Model behaviour
The sensitivity of the deep-sequencing approach also enabled investigators to detect additional trends, such as the widespread production of RNA molecules just 18 bases long near the start sites of genes2. Piero Carninci of Japan's Institute of Physical and Chemical Research (RIKEN), in Yokohama, used the FANTOM database, as well as others, to analyse RNA molecules produced from the remnants of retrotransposons, which once 'jumped' around the genome but are now thought to lie dormant. The team found that RNA is produced from these retrotransposon sequences in a highly regulated manner that varies among different cell stages and types3. Carninci says studies are under way to determine what function these transposons may have.
Aviv Regev of the Massachusetts Institute of Technology in Cambridge, says that for all its success, the FANTOM project highlights the challenges that systems biologists face as they strive to model the complex networks that control cell behaviour. "When you work with responses and circuits, the number of things you need to measure is substantially larger than when you simply sequence a genome," says Regev. "This is one cell line, one single signal, one time course." And yet the Japanese government, the sole funder of the project, has invested US$50 million in FANTOM over the past five years, according to organizer Yoshihide Hayashizaki of RIKEN's Yokohama Institute.
"We will not be successful as a community if we rely on things that require so many people and so much money," says Regev. Nevertheless, she remains optimistic that additional technological advances will speed things up. Future projects may take only weeks now that an analysis pipeline has been established, says Carninci. And although Bussemaker says that the FANTOM analysis was limited to transcription factors that have well-characterized DNA-binding sites in the genome — only a few hundred of more than 1,000 known human transcription factors — new techniques are allowing researchers to rapidly identify the DNA-binding preferences of the remaining set of proteins, he says.
"This paper is an important step on a path that's pretty long," says Regev. "On the other hand, I feel that we are on the brink of several of these important steps along this path."
References
The FANTOM Consortium and the Riken Omics Science Center Nature Genet. advance online publication doi:10.1038/ng.375 (2009).
Taft, R. J. et al. Nature Genet. advance online publication doi:10.1038/ng.312 (2009).
Faulkner, G. J. et al. Nature Genet. advance online publication doi:10.1038/ng.368 (2009).
Published online 19 April 2009 | Nature
