ABodyBuilder-ML \\ //
An Antibody Modelling Protocol \\ //
Built by OPIG ||
||
ABodyBuilder-ML uses:
o ANARCI to annotate the antibody sequence.
o SAbDab to select templates for the framework regions.
o ABangle to predict the VH-VL orientation.
o ABlooper to model the CDR regions.
o PEARS to build side chains.
Input parameters:
Job name = my_example_model
Job ID = my_example_model
Heavy sequence = VKLLEQSGAEVKKPGASVKVSCKASGYSFTSYGLHWVRQAPGQRLEWMGWISAGTGNTKYSQKFRGRVTFTRDTSATTAYMGLSSLRPEDTAVYYCARDPYGGGKSEFDYWGQGTLVTVSS
Light sequence = ELVMTQSPSSLSASVGDRVNIACRASQGISSALAWYQQKPGKAPRLLIYDASNLESGVPSRFSGSGSGTDFTLTISSLQPEDFAIYYCQQFNSYPLTFGGGTKVEIKRTV
Started modelling
Using anarci to annotate the target sequences
Heavy variable domain identified in heavy sequence
Species identified as human
VKLLEQSGAEVKKPGASVKVSCKASGYSFTSYGLHWVRQAPGQRLEWMGWISAGTGNTKYSQKFRGRVTFTRDTSATTAYMGLSSLRPEDTAVYYCARDPYGGGKSEFDYWGQGTLVTVSS
*************************************************************************************************************************
Trimming heavy sequence to the variable domain (*) only.
Kappa variable domain identified in light sequence
Species identified as human
ELVMTQSPSSLSASVGDRVNIACRASQGISSALAWYQQKPGKAPRLLIYDASNLESGVPSRFSGSGSGTDFTLTISSLQPEDFAIYYCQQFNSYPLTFGGGTKVEIKRTV
***********************************************************************************************************---
Trimming light sequence to the variable domain (*) only.
Writing the anarci annotation to file
Done. See my_example_model/alignments/target_numbering.txt
Choosing framework templates
Chose 1vge chain H with sequence identity over heavy chain framework of 1.00
Chose 1vge chain L with sequence identity over light chain framework of 1.00
Using 1vge H L as the orientation template
Using the angles HL=-59.72, HC1=72.89, HC2=114.93, LC1=118.83, LC2=83.20, dc=16.19 inherited from structure 1vge_H_L_0 with sequence identity 1.00
Assembling the framework template
Using ABangle to orientate the template chains
FREAD is gone! Now, ABlooper is running!
CDR goemetry refinement enabled 1
Decoy Diversity: {'H1': 0.46106514930725095, 'H2': 0.5595452308654785, 'H3': 1.2030884742736816, 'L1': 0.4156893253326416, 'L2': 0.34699125289916993, 'L3': 0.42172603607177733}
Skip sidechains False
Predicting side chains with PEARS
Running PEARS - 'force_fit' set to False
Detecting and building disulphide bridges first, as necessary...
Detecting and building unimodal rotamers...
Building remaining rotamers...
Precomputing energies of possible rotamers...
Running DEE...
Building an interaction graph...
Detecting and searching biconnected components...
Solving remaining clusters...
Filling in remaining gaps...
Now adding sidechains with PEARS!
Writing modelling report
Renumbering model files with the imgt numbering scheme
Finished modelling
Exiting with success code.