SABBAC v1.2: Structural Alphabet based protein Backbone Builder from Alpha Carbon trace.

Summary: SABBAC is an on-line service devoted to protein backbone reconstruction from alpha-carbon trace. It is based on the assembly of fragments issued from library of reduced size, resulting from the encoding of the protein trace in an HMM-derived structural alphabet [1, 2]. The assembly of the fragments is achieved by a greedy algorithm [3, 4], using an energy based scoring inspired from the OPEP force field [5]. Alpha-carbon coordinates remain unaffected. SABBAC simply positions the missing backbone atoms, no further refinement is performed. From our tests, SABBAC performs equal or better than other similar online approach and is robust to deviations on the alpha-carbon coordinates. It can be accessed at http://mobyle.rpbs.univ-paris-diderot.fr/cgi-bin/portal.py#forms::SABBAC.

Note: We are moving our servers. Please DO NOT use The previous link http://bioserv.rpbs.jussieu.fr/SABBAC.html any longer. .

v1.2 (September 2006): Systematic use of fast energy calculation. Now accepts large files. Slightly improved results.

• [ Top ]
• Fill the form
• Input a PDB File
• Number of models to build
• Position side chains
• Fast energy calculation
• CA trace encoding
• Results
• Generated PDB file
• Models energy analyse
• References
• [ Bottom ]

• Fill the form
• Input a PDB File

Whatever is the method you use to input protein atoms coordinates, your file has to respect the PDB file format. If your coordinates include others atoms than CA, only CA atoms will remain. Note that the current version of the service only support one PDB chain (one model) for each run. HETERO will be replaced by their standard equivalent following the PDB rules, and ALTERNATES residues will be discarded : only the first one is taken into account.
No warranty is provided if these instructions are not respected.

There are many ways to input PDB coordinates, choose between:

• Upload a PDB file with CA coordinates

Simply browse your hard drive to select file to upload, by clicking the Browse... button of the SABBAC form.

• Paste PDB file lines in the text area

Copy and paste coordinates in the text area looks like:

ATOM 1 CA ASP A 4 36.940 28.938 81.417 1.00 50.26 1MYS 157 ATOM 2 CA ALA A 5 40.439 28.042 80.415 1.00100.00 1MYS 158 ATOM 3 CA GLU A 6 41.502 28.674 83.982 1.00100.00 1MYS 159 ATOM 4 CA MET A 7 38.694 26.790 85.603 1.00 58.24 1MYS 160 ATOM 5 CA ALA A 8 39.405 23.157 85.446 1.00 58.82 1MYS 161 ATOM 6 CA ALA A 9 40.675 22.550 88.927 1.00 15.46 1MYS 162 ATOM 7 CA PHE A 10 37.125 23.062 90.107 1.00 35.52 1MYS 163 ATOM 8 CA GLY A 11 35.523 22.167 86.903 1.00 65.75 1MYS 164 ATOM 9 CA GLU A 12 32.499 20.562 88.478 1.00 15.88 1MYS 165 ATOM 10 CA ALA A 13 32.294 23.256 91.146 1.00 53.37 1MYS 166 ATOM 11 CA ALA A 14 32.939 26.034 88.661 1.00 4.59 1MYS 167 ATOM 12 CA PRO A 15 29.336 26.813 87.962 1.00 53.17 1MYS 168 ATOM 13 CA TYR A 16 28.869 27.607 91.673 1.00 48.49 1MYS 169 ATOM 14 CA LEU A 17 32.041 29.663 91.898 1.00 56.60 1MYS 170 ATOM 15 CA ARG A 18 32.126 31.434 88.582 1.00 11.30 1MYS 171 ATOM 16 CA LYS A 19 30.008 31.772 85.465 1.00 28.18 1MYS 172 ATOM 17 CA SER A 20 30.771 30.892 81.860 1.00 16.73 1MYS 173 ATOM 18 CA GLU A 21 32.746 33.539 80.050 1.00 4.84 1MYS 174 ATOM 19 CA LYS A 22 29.798 34.544 77.991 1.00 26.86 1MYS 175 ATOM 20 CA GLU A 23 27.220 34.305 80.683 1.00 60.50 1MYS 176 ATOM 21 CA ARG A 24 29.348 37.017 82.033 1.00 22.95 1MYS 177 ATOM 22 CA ILE A 25 29.777 38.728 78.724 1.00 31.88 1MYS 178 ATOM 23 CA GLU A 26 26.120 38.449 77.933 1.00 7.06 1MYS 179 ATOM 24 CA ALA A 27 25.434 40.144 81.207 1.00 32.19 1MYS 180 ATOM 25 CA GLN A 28 27.952 42.948 81.387 1.00 35.23 1MYS 181 ATOM 26 CA ASN A 29 27.507 43.875 77.746 1.00 48.92 1MYS 182 ATOM 27 CA LYS A 30 23.887 44.767 77.566 1.00 13.42 1MYS 183 ATOM 28 CA PRO A 31 21.751 47.864 77.196 1.00 68.34 1MYS 184 ATOM 29 CA PHE A 32 21.628 50.248 80.136 1.00 14.31 1MYS 185 ATOM 30 CA ASP A 33 19.880 53.586 80.813 1.00 31.57 1MYS 186 ATOM 31 CA ALA A 34 21.472 55.150 83.831 1.00 16.14 1MYS 187 ATOM 32 CA LYS A 35 19.429 58.314 84.131 1.00 49.20 1MYS 188 ATOM 33 CA SER A 36 16.258 56.290 84.327 1.00 25.79 1MYS 189 ATOM 34 CA SER A 37 17.104 52.925 85.853 1.00 26.51 1MYS 190 ATOM 35 CA VAL A 38 16.920 52.936 89.646 1.00 11.53 1MYS 191 ATOM 36 CA PHE A 39 16.338 51.121 92.981 1.00 49.50 1MYS 192 ATOM 37 CA VAL A 40 13.551 52.289 95.255 1.00 3.49 1MYS 193 ATOM 38 CA VAL A 41 11.897 52.071 98.632 1.00 47.42 1MYS 194
• Type a PDB Id in the text area in lower case

Type a PDB Id in lower case. In case of multiple chains file, type PDB Id in lower case, and chain character in upper case like:

1timA
• Paste a PDB URL instead of an id

Finaly, you can choose to input pdb file from an URL, like: (but it could be any url at any server, specifying a single PDB file)


http://www.pdb.org/pdb/downloadFile.do?fileFormat=pdb&compression=NO&structureId=1PGB
• Number of models to build

Since the heap size we use for greedy algorithm is 10, you can choose how many models (in the heap) do you want to have in the output PDB file.
Computing time is identical whatever the number of models you choose.

• Position side chains

By default, only SABBAC backbone rebuilt coordinates are provided, but you can have a complete protein structure by checking the corresponding option in the SABBAC form.
Side chain positionning is performed by SCIt (/cgi-bin/SCit) [6] with a new fast backbone dependant procedure.

• Fast energy calculation

This option is the main improvement of SABBAC v1.2. In fact, a new implementation of the energy function leads SABBAC to rebuilt protein structures much more faster than the previous one. Elsewhere, the results are quite better than those published for the first version [7].

PDB Code	Size	SABBAC v1.2	SABBAC v1 [7]	MaxSprout [8]	bb [9]
111M	154	0.21	0.29	0.42	0.91
1CTF	68	0.33	0.43	0.73	0.85
1IGD	61	0.37	0.36	0.44	0.68
1OMD	107	0.40	0.35	0.41	0.77
1SEMA	58	0.48	0.48	0.34	1.00
1TIMA	247	0.58	0.59	0.60	0.97
1UBQ	76	0.34	0.35	0.38	0.96
2CTS	437	0.38	0.40	0.45	0.86
2LYM	129	0.38	0.38	0.44	0.98
2MHR	118	0.44	0.50	0.54	0.88
2PCY	99	0.44	0.42	0.54	0.91
2WRP	104	0.29	0.30	0.42	0.87
4PTI	58	0.37	0.53	0.44	0.81
5NLL	138	0.38	0.37	0.46	0.85
1PXZA	346	0.48	0.55	0.54	0.96
1RKIA	101	0.56	0.58	0.44	0.88
1S7LA	177	0.34	0.29	0.36	0.86
1T70A	255	0.44	0.42	0.50	0.95
1TXOA	235	0.39	0.41	0.38	0.96
1V0ED	666	0.51	0.48	0.45	0.89
1V7BA	175	0.25	0.30	0.41	0.87
1VB5B	255	0.30	0.34	0.42	0.84
1VKCA	149	0.34	0.28	0.33	0.82
1VR4A	103	0.46	0.47	0.59	1.00
1VR9A	121	0.45	0.42	0.45	0.79
1WMHA	83	0.27	0.27	0.28	0.82
1WMIA	88	0.42	0.41	0.42	0.81
1WPBG	168	0.34	0.37	0.35	0.86
1X6JA	88	0.40	0.43	0.36	0.76
1XB9A	108	0.48	0.46	0.51	0.81
1XE0B	107	0.58	0.61	0.62	0.90
Mean	-	0.40	0.41	0.45	0.87
SD	-	0.09	0.09	0.10	0.07

Backbone reconstruction performance is evaluated compared with the native structure (backbone heavy atoms r.m.s.d). This test set is composed with a subset of those presented by SA Adcock [9], and of recent newcomers of the PDB.

• CA trace encoding

Act with care when you check this option. It allows you to force encoding even if ca trace is degenerated (incoherent CA_i-CA_i+1 distance, missing residues ...).

nota bene
In case of multiple chains file, please submit one chain per job or SABBAC cgi will refuse submission.

• Results
• Generated PDB file

In case of multiple models output, models are ranking from 0 to n with our energetic criteria, with best model at the first position. Depending of the option you selected in the SABBAC form, output pdb file contains backbone heavy atoms or complete protein structure (with side chains rebuilt by SCit backbone dependant procedure).

HEADER Generated By SABBAC @ RPBS
ATOM      1  N   THR     1      17.295  11.552   3.539  1.00 -3.41
ATOM      2  CA  THR     1      16.967  12.784   4.338  1.00 -3.41
ATOM      3  C   THR     1      15.686  12.734   5.160  1.00 -3.41
ATOM      4  O   THR     1      15.112  13.787   5.458  1.00 -3.41
ATOM      5  N   THR     2      15.215  11.569   5.603  1.00 -1.27
ATOM      6  CA  THR     2      13.856  11.469   6.066  1.00 -1.27
ATOM      7  C   THR     2      14.020  10.786   7.406  1.00 -1.27
ATOM      8  O   THR     2      14.794   9.818   7.465  1.00 -1.27
ATOM      9  N   CYS     3      13.380  11.246   8.492  1.00 -3.62
ATOM     10  CA  CYS     3      13.660  10.707   9.787  1.00 -3.62
ATOM     11  C   CYS     3      12.235  10.382  10.177  1.00 -3.62
ATOM     12  O   CYS     3      11.338  11.114   9.733  1.00 -3.62
...
ATOM    173  N   TYR    44      13.557  11.406  16.429  1.00 -2.82
ATOM    174  CA  TYR    44      13.257  10.745  15.081  1.00 -2.82
ATOM    175  C   TYR    44      14.231   9.645  14.675  1.00 -2.82
ATOM    176  O   TYR    44      14.625   9.555  13.509  1.00 -2.82
ATOM    177  N   ALA    45      14.594   8.787  15.622  1.00 -1.66
ATOM    178  CA  ALA    45      15.445   7.667  15.246  1.00 -1.66
ATOM    179  C   ALA    45      15.188   6.808  14.160  1.00 -1.66
ATOM    180  O   ALA    45      16.056   6.453  13.366  1.00 -1.66
ATOM    181  N   ASN    46      13.920   6.424  14.043  1.00 -1.37
ATOM    182  CA  ASN    46      13.512   5.395  12.878  1.00 -1.37
ATOM    183  C   ASN    46      12.232   5.625  12.878  1.00 -1.37
ATOM    184  O   ASN    46      12.236   4.793  11.972  1.00 -1.37
ATOM    185  OXT ASN    46      12.525   6.405  13.784  1.00 -1.37
TER
END
    

Note that the last column of the generated PDB file contains the energy of the residue i as presented in the XML result file.

• Models energy analyse

A per model and per residu analyse is provided to guide user through models selection. An XML energy analyse file is also provided. For each residu, the following energetic terms are detailled:



ESCSC	Side chain - Side chain interactions
EHB	Hydrogen bonds
EVdW	Van der Waals interactions
EPhi	Phi positive
ECA	CA valence
ESAt	Structural Alphabet letters transition
E(i)	Energy associated to the residue i


Note that those terms are pseudo-energy used to drive SABBAC during rebuilding. Compared to other energy functions, only some terms are considered.

• References

[1] Camproux AC, Tuffery P, Chevrolat JP, Boisvieux JF, Hazout S.
Hidden Markov model approach for identifying the modular framework of the protein backbone.
Protein Eng. 1999 Dec;12(12):1063-73.

[2] Camproux AC, Gautier R, Tuffery P.
A hidden markov model derived structural alphabet for proteins.
J Mol Biol. 2004 Jun 4;339(3):591-605.

[3] Tuffery P, Guyon F, Derreumaux P.
Improved greedy algorithm for protein structure reconstruction.
J Comput Chem. 2005 Apr 15;26(5):506-13.

[4] Tuffery P, Derreumaux P.
Dependency between consecutive local conformations helps assemble protein structures from secondary structures using Go potential and greedy algorithm.
Proteins. 2005 Dec 1;61(4):732-40.


[5] Sebastien Santini, Guanghong Wei, Normand Mousseau, and Philippe Derreumaux
Exploring the Folding Pathways of Proteins through Energy Landscape Sampling: Application to Alzheimer's Beta-Amyloid Peptide.
Internet Electron. J. Mol. Des. 2003, 2, 564-577


[6] Gautier R, Camproux AC, Tuffery P.
SCit: web tools for protein side chain conformation analysis.
Nucleic Acids Res. 2004 Jul 1;32(Web Server issue):W508-11.


[7] Maupetit J, Gautier R, Tuffery P.
SABBAC: online Structural Alphabet-based protein BackBone reconstruction from Alpha-Carbon trace.
Nucleic Acids Res. 2006 Jul 1;34(Web Server issue):W147-51.


[8] Holm L, Sander C.
Database algorithm for generating protein backbone and side-chain co-ordinates from a C alpha trace application to model building and detection of co-ordinate errors.
J Mol Biol. 1991 Mar 5;218(1):183-94.


[9] Adcock SA.
Peptide backbone reconstruction using dead-end elimination and a knowledge-based forcefield.
J Comput Chem. 2004 Jan 15;25(1):16-27.

RPBS - EBGM - Implemented by Maupetit & Tufféry - last-update: 02/10/06