Mean of Circular Quantities (MCQ) is a dissimilarity measure useful in comparison of 3D protein and/or RNA structures. It calculates an average difference between corresponding torsion angle values (rotations around bonds). More information can be found in:
Tomasz Zok, Mariusz Popenda, and Marta Szachniuk. 2014. "MCQ4Structures to Compute Similarity of Molecule Structures." Central European Journal of Operations Research 22(3):457–73.
Jakub Wiedemann, Tomasz Zok, Maciej Milostan, and Marta Szachniuk. 2017. "LCS-TA to Identify Similar Fragments in RNA 3D Structures." BMC Bioinformatics 18(1):456.
This project contains command-line tools to compare RNA 3D structures.
$ ./mcq-local
usage: mcq-local [-a <arg>] [-M <arg>] [-n <arg>] -t <arg> [-T <arg>]
-a,--angles <arg> Torsion angle types (separated by comma without
space), select from:
ALPHA,BETA,GAMMA,DELTA,EPSILON,ZETA,NU0,NU1,NU2
,NU3,NU4,ETA,THETA,ETA_PRIM,THETA_PRIM,CHI,PSEU
DOPHASE_PUCKER. Default is:
ALPHA,BETA,GAMMA,DELTA,EPSILON,ZETA,CHI
-M,--selection-model <arg> Selection query for 3D RNA model
-n,--names <arg> Model names to be saved in output files
(separated by comma without space)
-t,--target <arg> Path to PDB file of the native 3D RNA target
-T,--selection-target <arg> Selection query for native 3D RNA target
selection query syntax is CHAIN:RESNUM-ICODE:COUNT e.g. A:1:5 means a
selection of 5 residues starting with residue number 1 in chain Amcq-local is a tool which compares a set of 3D structures against a reference target. User can select a set of torsion angle types to be considered in the comparison process. Both the target and models can be aligned using selection queries.
The results are:
- Ranking of models with respect to MCQ value
- Detailed data about dissimilarity for each residue in every model
- Visualization of secondary structure with color-coded dissimilarities in each residue for every model
$ ./mcq-local -t /tmp/rp13/13_0_solution_4XW7_rpr.pdb -T A:1:45 -M A:1:45 /tmp/rp13/13_{A,B,D,X}*.pdb
Partial results are available in: /tmp/bbbfa598-4adf-42b5-b457-fd89601fbb59
13_Das_2_rpr.pdb 19.07
13_Bujnicki_5_rpr.pdb 19.27
13_Bujnicki_8_rpr.pdb 19.60
...
$ ls -l /tmp/bbbfa598-4adf-42b5-b457-fd89601fbb59/
drwxr-xr-x 2 tzok tzok 100 03-14 14:43 13_Bujnicki_5_rpr
drwxr-xr-x 2 tzok tzok 100 03-14 14:43 13_Bujnicki_8_rpr
drwxr-xr-x 2 tzok tzok 100 03-14 14:43 13_Das_2_rpr
-rw-r--r-- 1 tzok tzok 39147 03-14 14:43 table.csv
...Each directory with model name contains visualization and details about angular dissimilarity and MCQ for each residue.
A helper script colorbar.py can be used to produce colorbars from generated table.csv using ./colorbar.py -i table.csv.
$ ./mcq-lcs
usage: mcq-lcs [-M <arg>] -t <arg> [-T <arg>] -v <arg>
-M,--selection-model <arg> Selection query for 3D RNA model
-t,--target <arg> Path to PDB file of the native 3D RNA target
-T,--selection-target <arg> Selection query for native 3D RNA target
-v,--mcq-threshold-value <arg> Value of MCQ threshold in degrees
selection query syntax is CHAIN:RESNUM-ICODE:COUNT e.g. A:1:5 means a
selection of 5 residues starting with residue number 1 in chain Amcq-lcs is a tool to find the longest continuous segment which is dissimilar up to a cerain MCQ threshold.
The result is a set of information about found segments alignment: size, coverage and nucleotide positions in the chains.
$ ./mcq-lcs -t /tmp/rp13/13_0_solution_4XW7_rpr.pdb /tmp/rp13/13_Das_2_rpr.pdb -v 15
MCQ value: 15°
Number of residues: 19
Coverage: 31.667%
Target name: 13_0_solution_4XW7_rpr
First target residue: A.U7
Last target residue: A.G25
Model name: 13_Das_2_rpr
First model residue: A.A27
Last model residue: A.C45$ ./mcq-global
usage: mcq-global [-a <arg>] [-M <arg>]
-a,--angles <arg> Torsion angle types (separated by comma without
space), select from:
ALPHA,BETA,GAMMA,DELTA,EPSILON,ZETA,NU0,NU1,NU2,
NU3,NU4,ETA,THETA,ETA_PRIM,THETA_PRIM,CHI,PSEUDO
PHASE_PUCKER. Default is:
ALPHA,BETA,GAMMA,DELTA,EPSILON,ZETA,CHI
-M,--selection-model <arg> Selection query for 3D RNA model
selection query syntax is CHAIN:RESNUM-ICODE:COUNT e.g. A:1:5 means a
selection of 5 residues starting with residue number 1 in chain Amcq-global tool is useful for comparison of multiple 3D structures without a reference one. A single MCQ value is computed for each pair of 3D structures. The selection of torsion angle types to be considered in the comparison can be changed using --angles switch. By default, the alignment of structures is done automatically, but user can provide a selection query to compute just a fragment.
The results are:
- Distance matrix
- Clustering (k-medoids)
- Visualization
$ ./mcq-global /tmp/rp13/*.pdb
Comparing... 100% │████████████████████████████████│ 1081/1081 (0:00:09 / 0:00:00)
Results available in: /tmp/f3610dda-feec-460e-9cd4-ecf1c1b8466c
$ ls -l /tmp/f3610dda-feec-460e-9cd4-ecf1c1b8466c/
-rw-r--r-- 1 tzok tzok 5614 03-14 14:31 clustering.svg
-rw-r--r-- 1 tzok tzok 981 03-14 14:31 clustering.txt
-rw-r--r-- 1 tzok tzok 41926 03-14 14:31 matrix.csv
$ ./mcq-print
usage: mcq-print [-T <arg>]
-T,--selection-target <arg> Selection query for native 3D RNA target
selection query syntax is CHAIN:RESNUM-ICODE:COUNT e.g. A:1:5 means a
selection of 5 residues starting with residue number 1 in chain Amcq-print is a utility tool to print out values of torsion angles in a TSV format.
$ ./mcq-print /tmp/rp13/13_0_solution_4XW7_rpr.pdb
Chain ResNum iCode Name alpha beta gamma delta epsilon zeta chi
A 1 - G - -164.167 58.239 77.868 -153.691 -67.832 -170.106
A 2 - G -74.196 -178.465 57.158 79.161 -148.81 -69.172 -168.967
A 3 - G -63.156 173.711 54.802 76.679 -144.103 -80.547 -165.204
...