Northeast Structural Genomics Consortium (NESGC)

The NESG Consortium has invested significantly in technology development. This investment has paid off substantially in increased efficiency and quality of the structures we have produced, as documented in more than 60 scientific papers describing our technology development (refs 1-61 below), along with > 160 unique 3D protein structures and > 200 papers overall from the project to date. The key areas of technology development and integration have include: i) bioinformatics methods for clustering and prioritizing candidate proteins for biophysical analysis; ii) biotechnologies and robotic methods for high-throughput (htp) expression plasmid construction, expression screening, and protein production; iii) htp methods for analysis of ¡°foldedness¡± of expressed proteins by biophysical techniques; iv) cost efficient production of isotope or SeMet labeled protein samples for NMR or crystallographic analysis; v) experimental and theoretical methods for parsing large multi-domain proteins into domain encoding segments, and domain optimization; vi) robotic methods for htp protein solubility screening, crystallization screening, and crystal manipulation; vii) cryoprobe technologies and new GFT NMR pulse sequences providing rapid NMR collection; viii) automated computational methods for determining 3D protein structures from raw diffraction and/or NMR data; ix) computational methods for structure-based function prediction; x) integrated project databases to keep track of the reagents and information generated in a large-scale structural genomics effort; and xi) organized dissemination of the resulting expression plasmids, protein reagents, software, resonance assignments, functional annotations and 3D structures. There are also extensive efforts underway to develop new technologies for protein production, including the development of wheat germ cell-free and other eukaryote-based protein production systems. Retrospective analyses of the NESG databases provide important clues to construct design and conditions for improving expression and solublization success rates and crystallization. In the following sections, we briefly outline three highlights from these NESG technology development efforts.

 

At the beginning of PSI-1 in 2000, about 6 weeks of NMR instrument time per protein structure was considered to be a realistic estimate for ~1 mM protein samples with molecular weights up to 15 kDa. Since then, two major breakthroughs have dramatically reduced the NMR instrument time required to solve protein structures. First, cryogenic probes became available; these offer approximately a 3-fold increase in sensitivity in routine applications in biological NMR spectroscopy (14), and potentially an order-of-magnitude reduction in measurement times. However, it was recognized that conventional data collection in the ¡®sampling limited regime¡¯ (17) would prevent one from realizing this 10-fold reduction in NMR measurement time. In the sampling limited regime, a (large) fraction of instrument time is invested in sampling indirect dimensions of multidimensional NMR experiments, and not for achieving sufficient ¡®signal averaging¡¯. The ¡®NMR sampling problem¡¯ associated with this regime was effectively solved by introducing ¡®G-matrix FT NMR¡¯ spectroscopy (24, 36), which is rooted in ¡®Reduced-dimensionality NMR¡¯ spectroscopy (17). GFT NMR enables researchers to optimally adjust NMR measurement times to sensitivity requirements and allows them to take full advantage of highly sensitive cryogenic probes for high throughput NMR structure determination. Our progress in this area is demonstrated in references 14, 17, 19, 23, 29, 36, 41, 42, and 47 below.

The impact of GFT NMR with cryoprobes for HTP NMR structure determination has turned out to be truly spectacular. For example, recording a set of only 5 GFT NMR experiments on a 600 MHzb spectrometer equipped with a cryogenic probe (together with a simultaneous 15N, 13Cali, 13 Caroresolved [1H, 1H]-NOESY on a 750 MHz spectrometer equipped with a conventional probe), the Szyperski laboratory solved seven high-quality structures between 6/20/04 and 10/5/04; i.e. in about three months. The molecular weights ranged from 9 to 17 kDa for proteins with various different architectures: PfR14 (14 kDa, 7 days data collection time), ET95 (9 kDa, 2 days), XcR50 (9 kDa, 6 days), BcR68 (17 kDa, 7 days), MaR11 (11 kDa, 5 days), BhR29 (15 kDa, 4 days), SR215 (16 kDa, 5 days). The introduction of the GFT-based data collection methods allowed the Szyperski laboratory alone to produce some 14-protein NMR structures in Yr 4 of the NESG project. These results demonstrate the robustness and htp value of this new GFT NMR technology.

 

We have developed a fully integrated data analysis platform, which pulls together under a single Java interface the complete process of protein NMR structure determination and analysis, as well as archiving the raw NMR data (FID¡¯s) and intermediate results. SPINS (11, 48) integrates the Agnus and NMRPipe software for NMR spectral processing, the Auto Assign software for automated analysis of triple resonance and GFT NMR spectra, the Auto Structure, DYANA, and XPLOR software for automated NOESY analysis and structure calculation, and a wide range of tools for validating peak lists, resonance assignments and 3D structures. Auto Assign is very robust and has been used in the assignment more than 50 proteins by NESG members. Auto Structure has also been used to generate a large number of NESG structures, coming into routine use across the Consortium over the last year. Our progress in this area over the last four years is documented in references 8, 11, 14, 26, 30, 42, 43, 48, 52, 54, and 55 below.  The NESG has also developed fast and sensitive structure quality assessment tools, provided in the software AutoQF (55). AutoQF computes Recall, Precision, and F-measure scores (referred to as ¡°NMR RPF¡± scores), which measure of the goodness-of-fit of 3D structures with raw NOESY peak lists using methods from information retrieval statistics. These NMR RPF scores have been demonstrated to provide measures of correctness of the fold and accuracy of the structure (55). In addition, the introduction of NMR RPF scores in the NESG structure pipeline demonstrates the need to archive not only the NMR structures and constraint lists, but also the raw fid data, NOESY peaks lists from which constraint lists, and structures are derived. We anticipate that NMR RPF scores will have significant value to the field of structural biology, providing objective measures of the fit of experimental NMR structures to the data from which they are derived.

 

Protein crystallization remains a major bottleneck of hpt structure determination by X-ray crystallography. We have also made significant and substantial progress towards achieving the goal of fully automatic crystal image recognition. Automatic classification of protein images is critical in the development of high throughput crystallization screening and crystal sample mounting technologies. NESG has focused on the development of methods for the automatic classification of protein crystallization reactions based on image analysis of photomicrographs.  Given the great challenge and importance of this task, we have adopted a consensus approach for classification that includes two distinct recognition systems. Images are acquired by the DeTitta lab from a CCD camera under robotic control that captures the contents within each well of a 1536 matrix array. The images are taken at several times during an incubation period, scored manually in the DeTitta lab, and used for training of two independent image recognition systems. In the first recognition system, developed in the Jurisica lab (OCI), image analysis is performed in four stages: registration (locating the well within the image), segmentation (locating the drop within the well), feature extraction (computing 59 measures of drop texture), and classification (labeling the result as crystal, clear, other) (57). In the second method, under development in the Laine laboratory (Columbia), multiscale adaptive contrast enhancement is applied and an ellipsoid Hough transform is then performed to crop and isolate the drop within the well. A Laplacian pyramidal expansion is then performed and 16 features are extracted based on coefficient statistics. Finally, classification of the images is accomplished with a feed forward neural network. This work is described in references 2-5, 7, 21, 27, 33, 53, and 57 below.  Both image recognition systems described above have been validated and exhibit performance similar to human experts, approaching ~90% accuracy. Experimentally we have observed that humans and machines miss different types of crystals for distinct reasons. Through our consensus approach, we will best be able to minimize the chance of missing a crystal. As part of this effort, DeTitta lab is building, by manual classification, a database of (currently) over 1.5 million images that will provide ground truth in terms of a binary classifier: crystals and all other outcomes. By utilizing this extensive database with a consensus reasoning system, we hope to achieve better than- human performance during the next phase of development.

 

1. Bertone, P.; Kluger, Y.; Zheng, D.; Edwards, A.M.; Arrowsmith, C.H.; Montelione, G.T.; Gerstein, M. Nucleic Acids Research 2001, 29: 2884-2898. SPINE: An integrated tracking database and data mining approach for prioritizing feasible targets in high-throughput structural proteomics.

 

2. Jurisica, I.; Rogers, P.; Glasgow, J.; Collins, R.; Wolfley, J.; Luft, J.; DeTitta, G.T. IEEE Intelligent Systems Journal 2001, November/December, pp. 26-34. Improving objectivity and scalability in protein crystallization: Integrating image analysis with knowledge discovery.

 

3. Jurisica, I.; Rogers, P.; Glasgow, J.; Fortier, S.; Collins, R.; Wolfley, J.; Luft, J.; DeTitta, G.T. Methods in Macromolecular Crystallography 2001, L. Johnson and D. Turk (Eds.), Volume 325, NATO Science Series: Life Sciences, Kluwer Academic Press. High throughput macromolecular crystallization: An application of case-based reasoning and data mining.

 

4. Jurisica, I.; Rogers, P.; Glasgow, J.; Fortier, S.; Luft, J.; Bianca, D.; DeTitta, G.T. Thirteenth Annual Conference on Innovative Applications of Artificial Intelligence (IAAI-2001) 2001, Seattle, WA, p. 73-80. Image-feature extraction for protein crystallization: Integrating image analysis and case-based reasoning.

 

5. Jurisica, I.; Rogers, P.; Glasgow, J.; Fortier, S.; Luft, J.; Wolfley, J.; Bianca, M.; Weeks, D.; DeTitta, G.T. IBM Systems Journal 2001, 40: 394-409. Intelligent decision support for protein crystal growth.

 

6. Kulikowski, C. A.; Muchnik, I.; Yun, H. J.; Dayanik, A. A.; Zheng, D.; Song, Y.; Montelione, G.T. MedInfo 2001 ¨C 10th World Congress on Health and Medical Informatics 2001, 965-969. Protein structural domain parsing by consensus reasoning over multiple knowledge sources and methods.

 

7. Luft, J.; Wolfley, J.; Jurisica, I.; Glasgow, J.; Fortier, S.; DeTitta, G.T. J. Crys. Growth 2001, 232: 591-595. Macromolecular crystallization in a high throughput laboratory ¨C the search phase.

 

8. Moseley, H.N.B.; Monle¨®n, D.; Montelione, G.T. Methods in Enzymology 2001, 339: 91-108. Automatic determination of protein backbone resonance assignments from triple resonance NMR data.

 

9. Qian, J.; Stenger, B.; Wilson, C.A.; Lin, J.; Jansen, R.; Teichmann, S.A.; Park, J.; Krebs, W.G.; Yu, H.; Alexandrov, V.; Echols, N.; Gerstein, M. Nucleic Acids Res 2001, 29: 1750-1764. Parts List: A web-based system for dynamically ranking protein folds based on disparate attributes, including whole-genome expression and interaction information.

 

10. Wigley, W.C.; Stidham, R.D.; Smith, N.M.; Hunt, J.F.; Thomas, P.J. Nature Biotechnology. 2001, 19: 131-136. Protein solubility and folding monitored in vivo by structural complementation of a genetic marker protein.

 

11. Baran, M.; Moseley, H.N.B.; Sahota, G.; Montelione, G.T. J. Biomol. NMR 2002, 24: 113-121. SPINS: Standardized ProteIn NMR Storage. A data dictionary and object-oriented relational database for archiving protein NMR spectra.

 

12. Chen, J.; Acton, T.B.; Basu, S.K.; Montelione, G.T.; Inouye, M. J. Molec. Microbiol. Biotech. 2002, 4: 53-59. Enhancement of the solubility of proteins over expressed in Escherichicia coli by heat shock.

 

13. Mezouar, Y.; Allen, P.K. IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2002), Lausanne, Switzerland, Sep. 30 ¨C Oct. 4, 2002, pp. 1766-1771. Visual served micro-positioning for protein manipulation tasks.

 

14. Monle¨®n, D.; Colson, K.; Moseley, H.N.B.; Anklin, C.; Oswald, R.; Szyperski, T.; Montelione, G.T. J. Struct. Funct. Genomics 2002, 2: 93-101. Rapid analysis of protein backbone resonance assignments using cryogenic probes, a distributed Linux-based computing architecture and a set of semi-automated spectral analysis tools.

 

15. Nair, R.; Rost, B. Bioinformatics 2002, 2:78-86. Inferring sub-cellular localization through automated lexical analysis.

 

16. Sultan, M.; Wigle, D.; Cumbaa, C.; Maziarz, M.; Glasgow, J.; M.-S. Tsao; Jurisica, I. Bioinformatics 2002, 18: S111-S119. Binary tree-structured vector quantization approach to clustering and visualizing micro array data.

 

17. Szyperski, T.; Yeh, D.C.; Sukumaran, D.K.; Moseley, H.N.B.; Montelione, G.T. Proc. Natl. Acad. Sci. USA 2002, 99: 8009-8014. Reduced-dimensionality NMR spectroscopy for high throughput protein resonance assignment.

 

18. Liu, J.; Rost, B. PROTEINS: Struct. Funct. Bioinformatics 2004, 55: 678-688. CHOP Proteins into structural domain-like fragments

 

19. Xia, Y.; Arrowsmith, C.H.; Szyperski, T. J. Biomol. NMR 2002, 24: 41-50. Novel projected 4D triple resonance experiments for polypeptide chemical shift assignment.

 

20. Yee, A.; Chang, X.; Pineda-Lucena, A.; Wu, B.; Semesi, A.; Le, B.; Ramelot, T.; Lee, G.M.; Bhattacharya, S.; Gutierrez, P.; Denisov, A.; Lee, C.-H.; Cort, J.R.; Kozlov, G.; Liao, J.; Finak, G.; Chen, L.; Wishart, D.; Lee, W.; McIntosh, L.P.; Gehring, K.; Kennedy, M.A.; Edwards, A.M.; Arrowsmith, C.H. Proc. Natl. Acad. Sci. USA 2002, 99: 1825-30. An NMR approach for structural proteomics.

 

21. Cumbaa, C.; Lauricella, A.; Fehrman, N.; Veatch, C.; Collins, R.; Luft, J.; DeTitta, G.; Jurisica, I. . Acta Crystallographica D, 2003, 59:1619-1627. Automatic classification of sub-microlitre protein crystallization trials in 1536-well plates.

 

22. Goh, C.-S.; Lan, N.; Echols, N.; Douglas, S.; Milburn, D.; Bertone, P.; Xiao, R.; Ma, L.-C.; Zheng, D.; Wunderlich, Z.; Acton, T.; Montelione, G.T.; Gerstein, M. Nucleic Acids Res. 2003, 31: 2833-2838. SPINE 2: A system for collaborative structural proteomics within a federated database framework.

 

23. Hunt, J.F.; Deisenhofer, J. Acta Cryst. Section D 2003, 59: 214-224. Ping-pong cross validation in real-space: a method to increase the phasing power of a partial model without risk of phase bias.

 

24. Kim, S.; Szyperski, T. J. Am. Chem. Soc. 2003, 125: 1385-1393. GFT NMR, a new approach to rapidly obtain precise high dimensional NMR spectral information.

 

25. Kimber, M.; Vallee, F.; Houston, S.; Necakov, S.; Skarina, T.; Evdokimova, E.; Beasley, S.; Christendat, D.; Savchenko, A.; Arrowsmith, C.H.; Vedadi, M.; Gerstein, M.; Edwards, A.M. PROTEINS: Struct. Funct. Genetics 2003 51: 562-568. Data mining crystallization  databases: Knowledge-based approaches to optimize protein crystal screens.

 

26. Li, W.; Zhang, Y.; Kihara, D.; Huang, Y.J.; Zheng, D.; Montelione, G.T.; Kolinski, A.; Skolnick, J. PROTEINS: Struct. Funct. Genetics 2003, 53: 290-306. TOUCHSTONEX: Protein structure prediction using sparse NMR data.

 

27. Luft, J.R.; Collins, R.J.; Fehrman, N.A.; Lauricella, A.M.; Veatch, C.K.; DeTitta, G.T. J. Struct. Biol. 2003, 142: 170-179. A deliberate approach to screening for initial crystallization conditions of biological macromolecules.

 

28. Savchenko, A.; Yee, A.; Khachatryan, A.; Skarina, T.; Evdokimova, E.; Pavlova, M.; Semesi, A.; Northey, J.; Beasley, S.; Lan, N.; Das, R.; Gerstein, M.; Arrowsmith, C.H.; Edwards, A.M. PROTEINS: Struct. Funct. Genetics 2003, 50: 392-399. Strategies for structural proteomics of prokaryotes: Quantifying the advantages of studying orthologous proteins and of using both NMR and X-ray crystallography approaches.

 

29. Xia, Y.; Yee, A.; Arrowsmith, C.; Gao, X. J. Biomol. NMR 2003, 27:193-203. 1H(C) and 1H(N) total NOE correlations in a single 3D NMR experiment. 15N and 13C time-sharing in t1 and t2 dimensions for simultaneous data acquisition.

 

30. Zheng, D.; Huang, Y.J.; Moseley, H.N.B.; Xiao, R.; Aramini, J.; Swapna, G.V.T.; Montelione, G.T. Protein Science 2003, 12: 1232-1246. Automated protein fold determination using a minimal NMR constraint strategy.

 

31. Georgiev, A.; Allen, P.K.; Mezouar, Y. Microrobotics for Biomanipulation Workshop, IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2003), Las Vegas, NV, Oct. 27 ¨C 31, 2003. Microrobotic crystal mounting using Computer Vision.

 

32. Maxwell, K.L; Bona, D.; Lui, C-S ; Arrowsmith, C.H.; Edwards, A.M. Protein Science, 2003, 12: 2073-2080. Refolding out of guanidine hydrochloride is an effective approach for high throughput structural studies of small proteins.

 

33. Angelini, E.D.; Wang, Y. Laine, A.F. The Workshop on Genomic Signal Processing and Statistics (GENSIPS) Baltimore, MD. May 26-27, 2004. Classification of micro array genomic images with brushlet analysis and neural networks.

 

34. Alexandrov, V.; Gerstein, M. BMC Bioinformatics 2004, 5: 2. Using 3D Hidden Markov Models that explicitly represent spatial coordinates to model and compare protein structures.

 

35. Arshadi, N.; I. Jurisica. 7th European Conference on Case-Based Reasoning (ECCBR¡¯04), Advances in Case-Based Reasoning, LNAI 3155, Eds. Peter Funk and Pedro A. Gonzalez Carlo, Springer-Verlag, pp. 17-31, 2004. Maintaining CBR systems: A machine learning approach.

 

36. Atreya, H.S.; Szyperski, T. Proc. Natl. Acad. Sci. USA 2004, 101: 9642-9647. G-matrix Fourier transform NMR spectroscopy for complete protein resonance assignment.

 

37. Das, R.; Gerstein, M. PROTEINS: Struct. Funct. Bioinformatics 2004, 55: 455-463. A method using active-site sequence conservation to find functional shifts in protein families: application to the enzymes of central metabolism, leading to the identification of an anomalous isocitrate dehydrogenase in pathogens.

 

38. Everett, J.K.; Acton, T.B.; Montelione, G.T. J. Struct. Funct. Genomics 2004, 5: 13-21. Primer Prim¡¯r: A web based server for automated primer design.

 

39. Georgiev, A.; Allen, P.K.; Edstrom, W. IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2004), Sendai, Japan, Sep. 28 ¨C Oct. 4, 2004. Visually guided protein crystal manipulation using micromachined silicon tools.

 

40. Goh, C.-S.; Lan, N.; Douglas, S.; Wu, B.; Bertone, P.; Echols, N.; Smith, A.; Milburn, D.; Montelione, G.T.; Zhao, H.; Gerstein, M. J. Mol. Biol. 2004, 336: 115-130. Mining the structural genomics pipeline: Identification and analysis of protein properties that affect high throughput experimental analysis.

 

41. Kim, S.; Szyperski, T. J. Biomol. NMR. 2004, 28: 117-130. GFT NMR experiments for polypeptide backbone and 13C¦Â chemical shift assignments.

 

42. Moseley, H.N.B.; Riaz, N.; Aramini, J.M.; Montelione, G.T. J. Magn. Reson. 2004, 170: 263-277. A generalized approach to automated NMR peak list editing: Application to reduced dimensionality triple resonance spectra.

 

43. Moseley, H.N.B.; Sahota, G.; Montelione, G.T. J. Biomol. NMR 2004, 28: 341-355. Assignment validation software suite for the evaluation and presentation of protein resonance assignment data.

 

44. Przulj, N., Wigle, D., Jurisica, I. Bioinformatics 2004, 20: 340-348. Functional topology in a network of protein interactions.

 

45. Zheng, D.; Aramini, J.; Montelione, G.T. Protein Science 2004, 13: 549-554. Validation of helical tilt angles in the solution NMR structure of the Z domain of Staphylococcal protein A by combined analysis of residual dipolar and NOE data.

 

46. Acton, T.B.; Gunsalus, K.C.; Xiao, R.; Ma, L-C.; Aramini, J.M.; Baran, M.C.; Chiang, Y-W.; Climent, T.; Cooper, B.; Denissova, N.; Douglas, S.M; Everett, J.K.; Ho, C.K.; Macapagal, D.;Paranji, R.K.; Shastry, R.; Shih, L-Y.; Swapna, G.V.T.; Wilson, M.; Wu, M.; Gerstein, M.; Inouye, M.; Hunt, J.F.; Montelione, G.T. Methods in Enzymology 2004 (in press). Robotic cloning and protein production platform of the Northeast Structural Genomics Consortium.

 

47. Atreya, H.S.; Szyperski, T. Methods in Enzymology 2004 (in press) Rapid sampling of NMR data.

 

48. Baran, M.C.; Huang, Y.J.; Moseley, H.N.; Montelione, G.T. Chemical Reviews 2004, 104: 3451-3555. Automated analysis of protein NMR assignments and structures.

 

49. King, A.; Przulj, N.; Jurisica, I. Bioinformatics 2004 (in press) Protein complex prediction via cost-based clustering. Bioinformatics Advance Access published on June 4, 2004 Bioinformatics 2004; doi:10.1093/bioinformatics/bth351.

 

50. Qing, G; Ma, L.; Khorchid, A; Swapna, G. V. T.; Mal, T.K.; Takayama, M.M.; Xia, B.; Phadtare, S.; Ke, H.; Acton, T.; Montelione, G.T.; Ikura, M.; Inouye, M. Nature Biotechnology 2004, 22: 877-882. Cold-shock induced high-yield protein production in Escherichia coli.

 

51. Przulj, N.; Corneil, D.; Jurisica, I. Bioinformatics 2004 (in press). Modeling interactome: Scale-free or geometric? Bioinformatics Advance Access published on July 29, 2004 Bioinformatics 2004; doi:10.1093/bioinformatics/bth436.

 

52. Snyder, D.A. & Montelione, G.T. PROTEINS: Struct. Funct. Bioinformatics 2004 (in press). Automatic identification of core structural elements in conformational ensembles of proteins.

 

53. Wang, Y.; Angelini, E.; Mehra, M.; Kim, D.; Laine, A. IEEE Transactions on Biomedical Engineering 2004 (submitted). A microarray genomic image recognition system with Laplacian pyramid representations.

 

54. Huang, Y.J.; Moseley, H.N.B.; Baran, M.C.; Arrowsmith, C.; Powers, R.; Tejero, R.; Szyperski, T.; Montelione, G.T. Methods in Enzymology 2004 (in press). An integrated platform for automated analysis of protein NMR structures.

 

55. Huang, Y.J.; Powers, R.; Montelione, G.T. J. Am. Chem. Soc. 2004 (in press) Protein NMR Recall, Precision, and F-measure scores (RPF Scores): Structure quality assessment measures based on information retrieval statistics.

 

56. Brown, K.; Jurisica, I. Bioinformatics 2004 (submitted). Predicted human interaction database.

 

57. Cumbaa, C.A.; Jurisica, I. J. Struct. Funct. Genomics 2004 (submitted). Automatic classification and pattern discovery in high-throughput protein crystallization trials.

 

58. Choi, S.-G., Chiang, Y.-W., Chang, R., Ma, L.-C., Ho, C. K., Shastry, R., Acton, T., Montelione, G. T., and Anderson, S. J. Struct. Funct. Genomics 2004 (submitted). P. pastoris as an expression host for structural genomics: Construction of a defined test panel of genes integrated as single copies at the aox1 locus.

 

59. Georgiev, A.; Allen, P.K.; Song T.; Laine A.; Edstrom W.; Hunt J. IEEE International Conference on Robotics and Automation (ICRA 2005), Barcelona, Spain, April 18 ¨C 22. (submitted). Microrobotic streak seeding for protein crystal growth.

 

60. Liu, J.; Hegyi, H.; Acton, T.B.; Montelione, G.T.; Rost, B. PROTEINS: Struct. Funct. Bioinformatics 2004 56:188-200. Automatic target selection for structural genomics on eukaryotes

 

61. Wunderlich, Z.; Acton, T.B.; Liu, J.; Kornhaber, G.; Everett, J.; Carter, P.; Lan, N.; Echols, N.;Gerstein, M.; Rost, B.; Montelione, G.T. PROTEINS: Struct. Funct. Bioinformatics 2004, 56: 181-187. The protein target list of the Northeast Structural Genomics Consortium.