-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathCHANGELOG
78 lines (60 loc) · 3.98 KB
/
CHANGELOG
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
README for PIZSA webserver edition v2.2
This is meant to serve as a changelog document noting all the additions made to PIZSA webserver source code, version 2.2 onwards
The source code can be found in 'scripts/', which is further subdivided into three directories to serve three major functions -
'src_InterRes' - contains a subset of the cell_list algorithm scripts; computes the interface residues without running the prediction software
'src_main' - contains the cell_list algorithm as well as the prediction algorithm; computes all necessary parameters for PIZSA prediction
'src_mutanal' - contains code for alanine scanning predictions using PIZSA; can be used as a standalone program.
Additional functionalities are added as separated modules and older modules are deprecated when no longer needed.
Implementation of the additional modules is done in the 'predict_binding.py' script, beta-level changes will be implemented only in the 'predict_binding_interface' function
Finding clashes -
-- Modules added under 'src_main'
-- find_clashes.py
-- percent_clash.py
-- percent_clash_2.py
-- find_clashes.py takes in the interacting residue pairs defined by a certain distance threshold and checks for any potential clashes.
-- interacting atoms are classified as clashes if the distance between them is less than the sum of the van der Waal's radii of the constituent atoms
-- 'Clash severity' is defined as the difference between the distance threshold ('cut_off') used for defining interactions and the interatomic distance
-- percent_clash.py takes in the clashes defined by the previous script and the cifa values computed elsewhere penalizing the clashing residues
-- interacting residues classified as clashing are assigned a 'clash-severity' based on the fraction of interacting atoms involved in the clash
-- clash penalty is computed through a sigmoidal function that takes 'clash-severity' as input
-- DEPRECATED
-- percent_clash_2.py takes in the clashes defined by the previous script and the cifa values defined elsewhere penalizing the clashing residues
-- interacting atoms are classified as clashes according to the definition used in 'find_clashes.py'
-- 'new_cifa' for clashing residues is computed by dividing the 'old_cifa' by the 'clash-severity'
Scaling scores for clashing residues -
-- Modules added under 'src_main'
-- score_complex.py
-- score_complex_clash.py takes in a the maximum clashes between residue pairs on the interface and penalizes their scores
-- scale_dict output from percent_clash_2 returns the maximum atomic clash for each residue pair on the interface
-- for any residue pair on the interface that has atomic clashes, a positive constant ( 1 in the current iteration) is added to the raw score
-- the more numerous the clashes, the more unfavourable the raw score of the structure
RESULTS -
After percent_clash_2.py -
Raw scores as well as Normalized Raw Scores are still more favourable for the decoys than the native structures
After score_complex_clash.py -
PDB_name, Stable, Unstable, Native
t53, 0, 1355, Non-binder
t50, 1451, 0, Binder (2 Bind, 1 Non-bind)
t54, 0, 1387, Binder
t41, 0, 1172, Non-binder
t40, 0, 2128, Binder
t47, 0, 1051, Non-binder
t46, 0, 1555, Non-binder
t29, 0, 1972, Binder
t38, 0, 874, Binder
t39, 0, 1294, Binder
t36, 0, 309, Non-binder
t37, 1129, 371, Binder (all)
t35, 0, 499, Non-binder
t32, 0, 599, Binder
t30, 0, 1201, Binder
7 / 13 correct predictions for Native structures (t50 anq t37 are multimeric targets)
PIZSA webserver edition v2.4_beta
-- New modules : write_out_mm.py (Placeholder till I fix getting normalized z-scores for multimeric structures)
-- New functions/modifications:
-- 'score_complex_clash.py'
-- This script now calculates normalized raw scores for the randomized decoys
-- 'calc_zscore.py
-- Added 'calc_zscore_norm' to calculate the z_scores based on normalized raw scores
-- 'write_out.py'
-- write_out function now takes the normalized output values and prints it in the main output file