back to Table of Contents

14. Scanning mutagenesis

14.1 Overview

EGAD can be used for scanning mutagenesis on a template structure. The EGAD energy function is able to predict the relative affinity of >1500 mutants from over 30 proteins to within 1kcal/mol (Pokala and Handel 2005). A brief description of the thermodynamic cycle used for mutant stability prediction is shown in Figure 14.1.1. Figure 14.1.2 and Figure 14.1.3 show some of these data.

The scanning mutagenesis functionality is used with the same format as designs, except that "VARIABLE_POSITIONS" is replaced with "SCANNING_POSITIONS" in the inputfile. For example (examples/scanning_mutagenesis/scan_mut/gb1.scan_mut.input):
START
TARGET ../../template_structures/gb1.pdb
FORCEFIELD ../../energy_function/forcefield
LOOKUP_TABLE_DIRECTORY ../../lookup_tables/gb1.lookup_table
OUTPUT_PREFIX gb1.scan_mut
SCANNING_POSITIONS
43 W, E, K.
44 A, E.
45 Q, R, S.
46 all
47 polar
END

Instead of running a design, however, the residues specified for each position are substituted, one by one, their rotamers optimized, and their energies calculated. An important caveat (compared to the other jobtypes) is that no other residues will be allowed to move. The user must specify the positions whose rotamers are allowed to move by listing them in the SCANNING_POSITIONS section without an amino acid type; these will be allowed to move, with the template amino-acid identity.

The resulting output_prefix.scanning_mutagenesis file lists the following:
seq_pos wt_res mutant_res ∆∆Gfolded ∆clashes   ∆hbond_unsat   ∆∆Gtotal coresurfint

For example, examples/scanning_mutagenesis/scan_mut/gb1.scan_mut.scanning_mutagenesis file has:
seq_pos wt_res mutant_res ddG_folded clashes unsat_hbond ddG_total coresurfint
...
43 W W 0.000000 0 0 0.000000 s
44 T T 0.000000 0 0 0.000000 s
45 Y Y 0.000000 0 0 0.000000 s
46 D A 0.304587 0 0 1.522936 s
46 D D 0.000000 0 0 0.000000 s
46 D E -0.571090 0 0 -2.855454 s
46 D F -0.508344 0 1 12.172872 s
...

Asp (D) 46 (s = surface position) -> Ala (A) has a ∆∆G_folded = 0.304587 kcal/mol and a ∆∆G_total = 1.522936 (meanings of energy values discussed in output pdb file section). It does not form or relieve a clash, nor does it have any affect on the number of unsatisfied hydrogen-bonding atoms.

As discussed previously, predicted ∆∆Gs are unreliable if the difference if |∆clashes| > 2 (Pokala and Handel 2005).

Bu default, the files for the individual point mutants are deleted upon completion. To keep these files, include:
CLEAN_UP_FLAG 0 # default 1 removes intermediate files
in the inputfile.

By default, the coordinates of the point mutants are not saved to the pdb files, since the master only requires the energies and rotamers listed in the header of the output pdb file; writing the coordinates can take a significant amount of time. If the coordinates are to be kept, include:
OUTPUT_COORD_FLAG 1 # for scanning mutagenesis, default 0 does not save coordinates
in the inputfile

Note that since the non-structured reference states for cysteine have not been calculated, the energies for positions that contain cysteine in the template sequence are unreliable.

Scanning mutagenesis is readily parallelized:
EGAD.exe gb1.scan_mut.input parallel (parallel options here or in inputfile)
will parallelize the lookup table generation, as well as the rotamer optimization of the point-mutants.

14.2 alanine_scan and saturation_mutagenesis

The following is a shortcut for performing an alanine scan at all movable positions:
START
...
END
SCANNING_POSITIONS
alanine_scan
END

Saturation mutagenesis (allow every amino acid at all positions) may be defined in a similar manner:
START
...
END
SCANNING_POSITIONS
saturation_mutagenesis
END

For examples of these jobs, see examples/ala_scan/gb1.ala_scan.input,
examples/sat_mut/gb1.sat_mut.input, and their outputs for alanine scanning and saturation mutagenesis, respectively. For both of these jobs, for each mutant, rotamers at all other positions are allowed to move.

back to Table of Contents