Extract the PAE & pLDDT value of Alphafold ver. 2.0.0 from .pkl to .csv & .pdb file.
Alinment program of .pdb files were updated for Alphafold ver. 2.1.1
(align_mono.py, align_ptm.py, align_multi.py)
https://github.com/CYP152N1/AF2align
Introduction movie was upload on Youtube.
Japanese var. (plddt_align.py)
for Alphafold 2.1.1 (align_mono.py, align_ptm.py, align_multi.py)
https://youtu.be/sDsV7pr6T_A?t=358
"align_multi.py", "align_mono.py" and "align_ptm.py" were released.
This version was created for alphafold 2.1.1. to align .pdb files
This script can be parformed without any argument.
$ cd /path/to/output/directory
$ python align_mono.py
It is omitted that the inthersion of pLDDT at b-factor on ".pdb format", because pLDDT was already writtened in out put file of AF 2.1.
additional argument can be observed by
$ python align_mono.py -h
multimer version can be select alinment chain by
$ python align_multi.py -alc 2
The default value of -arc argument is 1 for the alignment of A chain.
These alignment program will be separated to another programs, because program content and name don't match.
"pae2csv.py", "pae2png.py", "plddt_align_ptm.py", and "ptm_docker.py" were released
[pae2csv.py] generates [p2c_pae_?.csv] & [p2x_plddt.csv].
[p2c_pae_?.csv] contains Predicted Aligned Error (PAE).
You can visualize using Excel like the figure shown below.
Thanks Dr. Yoshitaka Moriwaki (@Ag_smith) for showing how to export predicted aligned error in alphafold2.
To extract "PAE" into [result_model_?.pkl], we need to change [docker/run_docker.py]
[ptm_docker.py] creates run_docker_ptm.py from your run_docker.py.
model_names of [run_docker_ptm.py] was changed to "model_?ptm" from "model?"
$ cd /path/to/storage/
$ git clone https://github.com/CYP152N1/plddt2csv
$ cd /path/to/Alphafold2
$ python3 /path/to/Alphafold2/ptm_docker.py
$ python3 docker/run_docker_ptm.py --fasta_paths=/path/to/.fasta --max_template_date=2020-05-14
If you performe "run_docker_ptm.py" instead of "run_docker.py",
result_model_?_ptm.pkl posesses predicted_aligned_error value.
$ cd /path/to/output/of/Alphafold2/
$ python3 /path/to/storage/plddt2csv/pae2csv.py
or
$ python3 /path/to/storage/plddt2csv/plddt_align_ptm.py
If you perform "run_docker_ptm.py" instead of "run_docker.py"_ptm,
you cannot use plddt_align.py and plddt2csv.py
[pae2png.py] generates [p2c_pae_?.png] using matplotlib & seaborn.
You need to perform below comands.
$ pip3 install matplotlib
$ pip3 install seaborn
-m arguments was introduced in plddt_align.py & plddt_align_ptm.py.
If you don't perform amber_minimization,
or amber_minimization was clashed on first model
python3 plddt_align.py -m unrelaxed_model
or
python3 plddt_align.py -m unrelaxed_model -n 1
plddt_align.py was released
$ cd /path/to/output/of/Alphafold2
$ python3 /path/to/storage/plddt2csv/plddt_align.py
This program automatically generates [align_xxxx_?.pdb]s.
These .pdb files will be aligned each other.
These .pdb including pLDDT instead of b-factor.
-d arguments was introduced.
python3 plddt2csv.py -d 1-30_53-
or
python3 plddt_align.py -d 1-30_53-
avarage pLDDT will be caliculate without residue number from 1 to 30 and from 53 to C-terminus.
["current folder name"_?_pLDDT??.pdb] will be also generated. default value is ""
$ cd /path/to/storage/
$ git clone https://github.com/CYP152N1/plddt2csv
$ cd /path/to/output/of/Alphafold2
$ python3 /path/to/storage/plddt2csv/plddt2csv.py
or
$ python3 /path/to/storage/plddt2csv/plddt_align.py
This program generates [plddt.csv] and [relaxed_"current folder name"?.pdb] or [align"current folder name"_?.pdb]
form [features.pkl], [result_model_?.pkl] and [relaxed_model_?.pdb]
plddt.csv show plddt value of each residue of all relaxed models.
output pdb file contain plddt value in B factor region. High plddt value means high condidence of the output structure. You can check plddt put the colored by b-factor on the structure using model viewer such as pymol.
arguments can be chack by $ Python3 /path/to/storage/plddt2csv.py -h