SAP calculation procedure
By Naresh Chennamsetty
SAP (Spatial-Aggregation-Propensity) algorithm predicts the aggregation prone regions
of a protein (N. Chennamsetty et. al., PNAS, 106, 11937 (2009)). SAP also predicts protein
binding regions (N. Chennamsetty et. al., Proteins, 79, 888 (2011)). The procedure to
evaluate SAP for a protein using the scripts developed at MIT is detailed below. This
procedure gives a SAP value for a static structure (such as an X-ray structure). These
SAP values can be averaged over a simulation trajectory to get an average value for
simulation.
Requirements:
1. A computer with CHARMM simulation program installed (www.charmm.org)
The SAP scripts developed at MIT are written for the CHARMM
program. However, SAP algorithm can be re-written in any other script of
choice by the user.
2. Structure of the protein of interest (either from X-ray or homology modeling)
Summary of Steps:
1. Denote the PDB coordinate files for different chains within the protein as
‘chain1.pdb’, ‘chain2.pdb’…etc. Also set the ‘NUMchains’ parameter within the
file ‘inputs.str’ to the total number of chains.
2. Run the SAP script ‘sap.inp’ using the CHARMM program.
Detailed Steps:
1. Denote the PDB coordinate files for different (disjoint) chains within the protein
(obtained either from homology modeling or x-ray structure) as ‘chain1.pdb’,
‘chain2.pdb’…etc. Also change the ‘NUMchains’ parameter within ‘inputs.str’
script to the specified total number of chains. Any number of chains can be
supplied to the SAP script. For e.g., an antibody Fab fragment has two chains
(light and heavy chains), which should be denoted as ‘chain1.pdb’ and
‘chain2.pdb’, and ‘NUMchains’ set to 2. It is preferable that these files only
contain the coordinate information (i.e., lines starting with “ATOM...”) and no
other extra data or lines. If you have CYS-CYS (disulphide) bonds in your
protein and if you want to preserve these bonds, then enter the CYS residue
numbers manually in the file ‘cysbonds.str’ and change the parameter
‘DisulphideBond’ to ‘YES’. If you do not enter these CYS residues manually, the
program still runs and gives SAP values, but there will be a small error in SAP
values (near CYS residues).
2. Run the SAP script ‘sap.inp’ using the command line prompt such as
“charmm-executable < sap.inp > out.dat”. The file ‘find_sap’ has an example
execution run. The ‘sap.inp’ script calls all other scripts and inputs automatically
to calculate the final SAP score. The calculation itself is quite fast and takes less
than 5 minutes to find SAP values for a protein. If required, more accurate SAP
values can be obtained by averaging the SAP values from different protein
conformations obtained along a simulation trajectory. The several outputs
obtained after running the ‘sap.inp’ script are explained below.
Outputs:
1. ‘sap_score_each_atom.dat’ gives the SAP value at each atom (last column).
2. ‘sap_score_residue_average.dat’ gives the average SAP value for each residue.
3. ‘protein_sap_mapped.pdb’ gives the final structure (pdb) file where SAP values
are included in the beta field (second to last column). This pdb file can be viewed
in molecular visualization programs such as ‘VMD’ to see the protein surface
colored according to the SAP values (for e.g., high SAP regions in red color, and
low SAP regions in blue color). In VMD, select “Graphics > Representations >
Coloring method > Beta” to get the protein colored according to SAP values.
4. Note: By default, the SAP radius is set to R=10Å. This can be modified in the
file ‘inputs.str’ to any other SAP-radius of interest. Typically, setting R=10Å
shows broader hydrophobic patches (SAP at low resolution) whereas R=5 Å
shows more detailed hydrophobic patches (SAP at high resolution).