FAST - Help

Run FAST Online

Thanks to its excellent speed and reliability, the FAST algorithm makes structural alignment just a few clicks away. The FAST web server is easy to use. Just upload your structures (in PDB format) using the "Browse" buttons, and then click "Submit", the structural alignment will be delivered to your browser instantly. Depending on protein size, network traffic and server load, it may take up to 10 seconds. Following the links in the result page, you can download the alignment result (named align.out) and the rasmol script to visualize your alignment (named align.ras). The format and use of these files are described below. Result files are in the same format as generated by running FAST offline.

Run FAST Offline

It is more convenient to run FAST on your own computer. After you have downloaded the FAST executables for your computer, type command "fast --help" to get started. Below is the help message:

FAST - FAST Alignment and Search Tool, is a protein structure alignment algorithm developed by Jianhua Zhu at the Bioinformatics Program of Boston University. Please visit the FAST homepage at http://bu.wenglab.org/. USAGE: $ fast protein1 protein2 [-r rasmol-script] Description: %1 fn protein #1 %2 fn protein #2 -o|--out fn output file name [stdout] -r|--rasmol fn generate a rasmol script for the alignment -p|--path fn path to protein structures -h|--help display this message To visualize alignment under unix/linux: $ rasmol -script rasmol-script Email zhiping.weng@umassmed.edu for bug reports.

FAST takes two PDB files as input (here are two examples: d1bhga1 and d1cs6a3). FAST only reads the ATOM lines of CA's and ignores chain information. You may want to isolate the segment or chain of interest before running FAST. For example, the following command takes the CA lines of chain A from a PDB file:

$ grep '^ATOM.*CA.* A ' your.pdb

If the PDB files are in another directory, you may specify the directory using the option "-p <path-to-pdb>". The alignment will be sent to the standard output unless the user specifies an output file using "-o <output-file-name>". Use option "-r <rasmol-script-name>" to visualize the alignment (see below). FAST uses a single set of thoroughly tested parameters and no case-to-case adjusting is necessary. Below is the backend command line when you run FAST online.

$ fast protein1 protein2 -o align.out -r align.ras

Interpreting the Alignment

Let's take the alignment of d1bhga1 and d1cs6a3 (both are taken from SCOP version 1.61) as an example:




FAST ALIGNMENT: d1bhga1 d1cs6a3

L=73 SX=5.649e+02 SN=5.835e+00 L1=103 L2=91 RMSD=3.295



 1:	----TYIDDIT---VTTSVEQDS-GLVNYQISVKGSNLFKLEVRLLDAENKVVAN--GTG

 2:	RQYAPSIKAKFPADTYALT----GQMVTLECFAFGNPVPQIKWRKLDGSQTSK--WLSSE



 1:	TQGQLKVPGVSLWWPYLMHERPAYL-------YSLEVQLTAQTSLGP-VSDFYTLPVGIR

 2:	PLLHIQ-------------------NVDFEDEGTYECEAENI-----KGRDTYQGRIIIH



 1:	T*

 2:	A*



  226  THR T  P PRO   213 

  227  TYR Y  S SER   214 

  228  ILE I  I ILE   215 

  229  ASP D  K LYS   216 

  230  ASP D  A ALA   217 

  231  ILE I  K LYS   218 

  232  THR T  F PHE   219 

  233  VAL V  T THR   223 

  234  THR T  Y TYR   224 

  235  THR T  A ALA   225 

  236  SER S  L LEU   226 

  237  VAL V  T THR   227 

  242  GLY G  Q GLN   229 

  243  LEU L  M MET   230 

  244  VAL V  V VAL   231 

  245  ASN N  T THR   232 

  246  TYR Y  L LEU   233 

  247  GLN Q  E GLU   234 

  248  ILE I  C CYS   235 

  249  SER S  F PHE   236 

  250  VAL V  A ALA   237 

  251  LYS K  F PHE   238 

  252  GLY G  G GLY   239 

  253  SER S  N ASN   240 

  254  ASN N  P PRO   241 

  255  LEU L  V VAL   242 

  256  PHE F  P PRO   243 

  257  LYS K  Q GLN   244 

  258  LEU L  I ILE   245 

  259  GLU E  K LYS   246 

  260  VAL V  W TRP   247 

  261  ARG R  R ARG   248 

  262  LEU L  K LYS   249 

  263  LEU L  L LEU   250 

  264  ASP D  D ASP   251 

  265  ALA A  G GLY   252 

  266  GLU E  S SER   253 

  267  ASN N  Q GLN   254 

  268  LYS K  T THR   255 

  269  VAL V  S SER   256 

  270  VAL V  K LYS   257 

  273  GLY G  S SER   260 

  274  THR T  S SER   261 

  275  GLY G  E GLU   262 

  276  THR T  P PRO   263 

  277  GLN Q  L LEU   264 

  278  GLY G  L LEU   265 

  279  GLN Q  H HIS   266 

  280  LEU L  I ILE   267 

  281  LYS K  Q GLN   268 

  301  TYR Y  G GLY   276 

  302  SER S  T THR   277 

  303  LEU L  Y TYR   278 

  304  GLU E  E GLU   279 

  305  VAL V  C CYS   280 

  306  GLN Q  E GLU   281 

  307  LEU L  A ALA   282 

  308  THR T  E GLU   283 

  309  ALA A  N ASN   284 

  310  GLN Q  I ILE   285 

  316  VAL V  G GLY   287 

  317  SER S  R ARG   288 

  318  ASP D  D ASP   289 

  319  PHE F  T THR   290 

  320  TYR Y  Y TYR   291 

  321  THR T  Q GLN   292 

  322  LEU L  G GLY   293 

  323  PRO P  R ARG   294 

  324  VAL V  I ILE   295 

  325  GLY G  I ILE   296 

  326  ILE I  I ILE   297 

  327  ARG R  H HIS   298 

  328  THR T  A ALA   299

The output file contains three components from top to bottom:

1. A header line beginning with "FAST ALIGNMENT:" to indicate the pair of protein structures aligned. The next line contains a basic description of the alignment. Each field is in the format of "name=value", explained in the following table

L	Length, or the number of residue pairs aligned
SX	The raw score, or total similarity
SN	Normalized score
L1	Number of residues in protein #1
L2	Number of residues in protein #2
RMSD	Root-mean-square distance after superposed

Always use the normalized score to measure the significance of an alignment. Usually, SN greater than 1.5 indicates significant structural similarity. The raw score depends heavily on the sizes of the protein structures. As a result, larger proteins tend to produce bigger raw scores by chance.

2. The next block displays the structure alignment in a sequence alignment format. Lines beginning with "1:" are for protein #1 and "2:" for protein #2. The alignment ends with a pair of *'s.

3. The last block contains lines of aligned residues, one line per pair. For instance, the first line in our example means that "226 THR" in d1bhga1 is paired to "213 PRO" in d1cs6a3.

Visualize the Alignment

To visualize aligned structures, generate a rasmol script using the option "-r <rasmol-script-name>". When running FAST through the web server, just click the link to save the script. Rasmol is a powerful software to display molecular structures. It is freely available here.

Once you have installed Rasmol and have gotten a rasmol script, say align.ras, for your alignment, simply type "rasmol -script align.ras" to view it. The first protein is colored red and the second cyan. The aligned backbone segments are thickened. On Windows systems, the rasmol program is named "rw32b2a.exe". In order to invoke rasmol with the option -script, please type "rw32b2a -script your.script" either in a DOS window or by clicking Start|Run.