Sir

We would like to respond to comments made in your News story “'Big science' protein project under fire” (Nature 443, 382; 2006) about Japan's Protein 3000 Project. This government project funds nine centres, including the RIKEN Structural Genomics/Proteomics Initiative, consisting of the RIKEN Genomic Sciences Center's Protein Research Group (http://protein.gsc.riken.jp) and the RIKEN SPring-8 Center.

First, we do not agree that the information gained is “of limited use” or, as one researcher is quoted as saying “a lot of it is junk”. RIKEN has made major contributions to structures and structural models of functionally important proteins. It is expected to have determined 2,500 new structures by spring 2007, or 5% of the entire Protein Data Bank (PDB). Nearly half of those determined so far were obtained using NMR, and consist of functional domains from biologically important (including disease-related, signal transduction and nucleic-acid-binding) human/mouse proteins. Multiple structures from each family were analysed to understand binding specificities. About 70% of all NMR structures of human/mouse proteins deposited in the PDB in 2005 were from RIKEN.

The Protein 3000 project contributes significantly to the goals of the International Structural Genomics Organization (http://www.isgo.org) by providing large numbers of templates that can be used to model other members of the protein families. On average, each NMR structure from RIKEN has contributed to about 300 new homology models, and each X-ray structure to about 200 new models at a level of 30% sequence identity. Quality assessment measures for RIKEN structures are similar to those for structures deposited in the PDB in 2000–2006 by traditional structural biology groups, according to Gaetano Montelione of the Northeast Structural Genomics Consortium (personal communication).

Second, we very strongly disagree with the comment that “A centre of that size should contribute to methodology, but there has been nothing.” RIKEN has made seminal contributions to the development of methodologies and technologies. RIKEN has pioneered cell-free protein synthesis on a production scale, and developed technologies for extensive sample optimization process using the cell-free method. These technologies have been indispensable in solving the structures of many difficult proteins. More than 1,000 NMR structures have been determined from protein samples synthesized by RIKEN's implementation of the cell-free method. Additionally, at RIKEN, N. Kobayashi has developed the KUJIRA software for spectral analysis and P. Güntert has developed the CYANA software for automated protein structure analysis. RIKEN will be opening the NMR facility, together with these important technologies, to external scientists in 2007.