1. Zhang, T., He, Y., Wang, J.W., Wu, L.J., Zheng, C.D., Hao, Q., Gu, Y.X. and Fan, H.F. (2012) Institute of Physics, Chinese Academy of Sciences Beijing, P. R. China

2. OASIS (One-wavelength Anomalous scattering or Single Isomorphous Substitution) A direct-method phase extension/refinement program based on a partial structure, which can be just the heavy-atom sub- structure or it can be from fragments to a nearly complete model SAD/SIR information will be beneficial but not compulsory IPCAS (Iterative Protein Crystal-structure Automatic Solution ) A pipeline from reflection files (*.sca or *.mtz) to a structure model (*.pdb and *.mtz) Phasing: OASIS, ShelxC/D/E, PHASER Density modification: DM, RESOLVE Model building/refinement: ARP/wARP, PHENIX.AutoBuild, Buccaneer

4. 结构测定周期过去现在 青霉素～ 3年～ 3年 1～2天1～2天 维生素 B 12 ～ 8年～ 8年 1～2天1～2天 胰岛素～ 14 年～1周～1周 肌红 / 血红蛋白 > 20 年～1周～1周

7. Ab initio direct methods: (M.M. Woolfson’s team, 1988 ~) SAYTAN (M.M. Woolfson’s team, 1988 ~) (H.A. Hauptman’s team, 1993 ~) SnB (H.A. Hauptman’s team, 1993 ~) (G.M. Sheldrick’s team, 1994 ~) SHELXD (G.M. Sheldrick’s team, 1994 ~) (M.M. Woolfson’s team, 2000 ~) ACORN (M.M. Woolfson’s team, 2000 ~) Direct-method phase extension/refinement: (Combining direct methods with other protein phasing methods) Fan, H.F. Acta Phys. Sin. 21, 1114-1118 (1965) Fan, H.F. Acta Phys. Sin. 21, 1114-1118 (1965) Sayre, D. Acta Cryst. A28, 210-212 (1972) Sayre, D. Acta Cryst. A28, 210-212 (1972) Hauptman, H.A. Acta Cryst. A38, 632-641; 289-294 (1982) Hauptman, H.A. Acta Cryst. A38, 632-641; 289-294 (1982) Giacovazzo, C. Acta Cryst. A39, 585-592 (1983) Giacovazzo, C. Acta Cryst. A39, 585-592 (1983) Fan, H.F. & Gu, Y.X. Acta Cryst. A41, 280-284 (1985) Fan, H.F. & Gu, Y.X. Acta Cryst. A41, 280-284 (1985) Hao, Q., Gu, Y.X., Zheng, C.D. & Fan, H.F. OASIS Hao, Q., Gu, Y.X., Zheng, C.D. & Fan, H.F. J. Appl. Cryst. 33, 980-981 (2000) J. Appl. Cryst. 33, 980-981 (2000)

8. SAD/SIR iteration Iterative phasing/model building based on the heavy- atom substructure bearing SAD/SIR information MR iteration Iterative phasing/model building based on a structure model, which can be from fragments to a nearly complete model without SAD/SIR information Phase extension Iterative phasing/model building based on a subset of known phases SAD/SIR iteration Iterative phasing/model building based on the heavy- atom substructure bearing SAD/SIR information MR iteration Iterative phasing/model building based on a structure model, which can be from fragments to a nearly complete model without SAD/SIR information Phase extension Iterative phasing/model building based on a subset of known phases

9. ‘SAD iteration’ in IPCAS Acta Cryst. D60, 1991-1996 (2004) Acta Cryst. D62, 883-890 (2006) OASIS DM RESOLVE AutoBuild (Phenix) ARP/wARP/RefMac Buccaneer/RefMac Direct-method SAD phasing SAD data Ok? End Partial structure Density modification Model building and refinement Yes No

10. Bimodal distribution from SAD The phase of F” Phase information available in SAD given the known heavy-atom substructure Phase information available in SAD given the known heavy-atom substructure Peaked at 1 Sim distribution 2 Cochran distribution Peaked at any where from 0 to 2  3 Direct-method SAD phasing

11. Acta Cryst. A41, 280-284 (1985) Cochran distribution combined with bimodal SAD distribution Sim distribution Bimodal SAD distribution Direct-method SAD phasing

12. X-rays  =0.92Å 2.9Å Ba-SAD data Redundancy=3.9 SCIENCE 306, 104 (2004) X-rays =0.9789Å 3.3Å Se-SAD data Redundancy=3.3 Nat. Struct. Mol. Biol. 13, 589 (2006) X-rays Cu-K  2.1Å S-SAD data Acta Cryst. D61, 1533 (2005) Chinese Phys. B 17, 1 (2008) =0.9796Å; 2.3Å Se-SAD data Auto-Rickshaw built a 57% model before and a 89% model after using OASIS J. Mol. Biol. 383, 49–61 (2008) X-rays Cr-K  2.3Å S-SAD data 213 residues with 2 S atoms in the ASU 。

13. Direct method phase refinement Ok? End Partial structure Density modification Model building and refinement ‘MR iteration’ in IPCAS OASIS DM RESOLVE AutoBuild (Phenix) ARP/wARP/RefMac Buccaneer/RefMac Yes No

14. Partial structure Phasing part of ‘MR iteration’  ” h is a set of ‘reference phases’ calculated each cycle from randomly selected 5% of the current structure model; |  h | is defined as the absolute difference between the current phase and  ” h Acta Cryst. D63, 793-799 (2007)

15. 46 residues 13 with side chains MR model MR model Cycle 2 ARP/wARP-DM iteration Cycle 1 Final model Final model 215 residues Cycle 1 Cycle 3 OASIS - DM-ARP/wARP iteration Cycle 7 Cycle 5 201 residues all with side chains E7_C–Im7_C MR-model completion with E7_C–Im7_C data (1ujz)

16. Model resulting from SAD phasing MR iteration Oasis - DM - Buccaneer - Refmac Final 1086 residues in the AU MR iteration based on a Phenix SAD model Tom70p (2gw1) 3.3Å Se-SAD data, redundancy = 3.3 Residues build1074906 Residues assigned1017207 R / R free 0.285 / 0.4030.48 / 0.59

17. Model resulting from MRSAD phasing MR iteration Oasis - DM - Buccaneer - Refmac Final 1356 residues in the AU MR iteration based on an Auto-Rickshaw MRSAD model TM1782 (1vkn) 1.8Å Se-SAD data, redundancy = 6.1 Residues build13691234 Residues assigned13531068 R / R free 0.230 / 0.3000.261/0.342

18. A latest application of MR model completion MBP-NDM1 Data (2.25Å resolution) was used (Zhang and Hao, The FASEB Journal, 2011) Target structure Without OASIS With OASIS

19. As defined in MR iteration Direct-method phase extension Input known phases to be kept fixed until the resultant model becomes big enough Initial model Phenix.Autobuild running with ‘quick’ and ‘helixes & strands only’

20. AutoBuild quick + 20 cycles of Oasis-DM-Buccaneer iteration Starting with 4.5Å known phases and 2.8Å F obs Final model AutoBuild quick + 20 cycles of DM-Buccaneer iteration 1h3i Phase extension from 4.5 to 2.8Å An example of direct-method phase extension

21. SIRAS data DM, AutoBuild Partial model MR completion Native data SIRAS phasing SAD phasing End Derivative 2 1 Treatment of SIRAS data 1 DM, AutoBuild Partial model MR completion SIRAS phasing End SIRAS data SAD phasing DM, AutoBuild Partial model MR completion Native data End Derivative 2

22. SIRAS resultant model improved by the ‘MR iteration’ of OASIS-IPCAS SIRAS model from 2.13Å native & 2.65Å Hg-derivative data Model after ‘MR iteration’ based on the SIRAS model and 2.13Å native data Final model 1

23. Better results from SAD phasing of the derivative followed by direct-method aided model completion of the native 2 Direct-method SAD phasing of 2.65Å derivative data Direct-method aided model completion applied to 2.13Å native data Final model

24. 4.0Å model from direct-method SAD phasing 2.13Å model after direct-method phase extension 2.13Å model after direct-method aided model completion SIRAS data with derivative at 4Å resolution – a simulation: Hg-derivative data truncated to 4.0Å Final Model