Backbone Matching (BM)

streamlined protein-protein interface loop mimicry strategy
a novel strategy of loop mimicry to discover mimics by overlaying cyclo-organopeptides on loop targets and find hits by assessing goodness of fit.

Usage

requirements

system: Linux (modified scripts for Windows, tested by Parn, are also uploaded)
language: Python
packages: Pymol, Pandas, Numpy

dataset

scaffolds: SMILES strings of 86 rigid and cheap organic scaffolds mentioned in the paper.
cyclo-organopeptides: SMILES strings of 602 cyclo-organopeptides, comprised of Ala and organic fragments, categorized by sequence length.
conformers: include a link to access generated backbone conformers of 602 cyclo-organopeptides.

scripts

prerequisites: prepare the following folders under the directory of scripts.
  1) loop_target: put loop structural files here
  2) conformers_for_matching: put unzipped conformer files of 602 cyclo-organopeptides here
      *important: please place the unzipped conformer files in the following order: conformers_for_matching/(length, 4-10)/86 cyclo-organopeptides for each length/corresponding conformers;
       here is an example of a conformer path: conformers_for_matching/4/4R0/4_0_0.mol2
  3) align_result: empty folder, to store the RMSD values from the alignment calculation
  4) mimic_result: empty folder, to store best hits from the alignment calculation

precise-alignment: assume a hot loop of length k, overlay this loop with cyclo-{-(Ala)k-organo-}.
prerequisite: cut the loop structure file (in PDB format) to the desired fragment
input: python precise_alignment.py 'loop-name' 'number of hits to store' 'length of input loop'
example input: python precise_alignment.py 'uPA' 50 8
output: RMSD values in folder 'align_result', overlays of best hits with the loop fragment in the same folder as the script in Pymol pse format

auto-slicing-alignment: slice a long loop to fragments with 4 - 10 amino acids, then precisely overlay each fragment with cyclo-{-(Ala)n-organo-}, n = length of the fragment.
prerequisite: extract the long loop from the original PDB file, then save it in pdb format
input: python auto-slicing-alignment.py 'loop-name' 'number of hits to store' 'length of input loop'
example input: python auto-slicing-alignment.py 'uPA' 50 10
output: RMSD values in folder 'align_result', overlays of best hits with different loop fragments in the same folder as the script in Pymol pse format

virtual screening: screen a library of loops using precise alignment.
This is the script used to screen 1398 Kritzer hot loops.  The results are in a zip file shared in the published paper.

illustrative output

virtual screening output: best cyclo-organopeptide mimics for 1398 hot loops 
another two illustrative outputs for auto-slicing and precise-alignment:
3BT1_uPA_scan.pse:  a graphical presentation of a thorough scan over uPA hot loop (20-31) by auto-slicing alignment
uPA_23_30_precise_scan.pse:  a graphical presentation of precise scan over uPA segment (23-30, including all 5 hot spots) by precise alignment

Citation

Mi, Tianxiong, Siriwibool, Siriwalee, Burgess, Kevin, Angew. Chem. Int. Ed. 2023, e202307092, https://onlinelibrary.wiley.com/doi/abs/10.1002/anie.202307092