AFTME is an alignment-free method for the automatic mapping of 3D motifs to different FGs of a specific ligand through two-dimensional clustering. This repository contains the source codes, datasets and results for large-scale identification of FG-based protein-ligand binding motifs.

Any questions or requests, please contact us: [email protected] or [email protected]

• AFTME is written in Python, Python>=3.6 is needed.

• Python Packages: Biopython, pandas, seaborn, Matplotlib, SciPy, NumPy

• -i/--inputfile: the 2 paths in order:

1. the absolute path of an excel file recording all 233 ligands information.

2. the absolute path of a folder containing txt files recording ligand nonredundant pdbID for each of 233 ligands.

• -o/--outfile:
  the absolute path of a folder to store distance matrix excel files created for each of 233 ligands.

python DistMat.py -i "./Datasets/Ligand_information.xlsx" "./Datasets/NonredundantLigID_folder" -o "DistanceMatrix"

• -i/--inputfile: the 2 paths in order:

1. the absolute path of an excel file recording all 233 ligands information

2. the absolute path of a folder placing distance matrix excel files created for each of 233 ligands

• -o/--outfile: the 3 paths in order:

1. the absolute path of a folder for saving clustering 233 heatmaps of the 233 ligands

2. the absolute path of a folder for saving clustering result texts recording atoms of function group and binding motifs

3. the absolute path of a folder for saving excel file recording the first stage scores of 181 ligands with 2 or more funtion groups out of 233 ligands.

python AFTME.py motifGen -i "./Datasets/Ligand_information.xlsx" "DistanceMatrix" -o "./Motif_all/heatmaps" "./Motif_all/statistic" "scorefolder"

• -i/--inputfile: the 3 paths in order:

1. the absolute path of an excel file recording inherent atom composition for each manualy predefined function group of each of 233 ligands.

2. the absolute path of a folder placing txt files recording clustered atoms of function group and binding motif for each of 233 ligands, which is genetated from motifGen.

3. the absolute path of a folder placing excel files recording 1st stage scores, which is generated from motifGen.(note: the folder is also used for saving the second stage scores of outfile)

• -o/--outfile:
  the absolute path of a folder for saving an excel file recording assignment between artificially defined function groups and clustered function groups of each of 233 ligands.

python AFTME.py assign_score -i "atom_233.xlsx" "statisfolder" "scorefolder" -o "assignfolder"

1. Motif_all: This folder contains the heatmaps, statistics results and figures for all the FG-binding motifs.

• Heatmaps: Heatmaps for all the ligands generated with AFTME

• motif_analyse: Bar plots, pie plots and venn plots to visualize different FG-binding motifs

• statistics: Dict files recording the motif informations for all the ligands

2. Datasets: This folder contains the datasets used for large-scale FG-motif identification.

• NonredundantLigID: The .txt files containing all the PDB codes used for motif extraction for each ligand

• Ligand_information.xlsx: The excel file recording the information: name, functional groups, 2d figures etc. for all the ligands involved in our analysis.

• PDBCodes.xlsx: The excel file recording all the PDB codes used for analysis

3. DistanceMatrix: This folder contains the excel files recording all the distance matrix generated for all the ligands.

4. Systematic_analysis: This folder contains the files and source codes to reproduce the results for systematic analysis.

• Cluster_file: Dict file recording the motifs in each cluster.

• FG_binding_motif_all.csv: Table recroding the information: ID, atom properties, aa compositions, etc. for all the identified motifs.

• motif_assign.xlsx: Table recording the clustering assignments for each FG in the analysis.

• Motif_clustering.ipynb: Source codes for the motif clustering: feature matrix construction, "elbow" plot, t-SNE plot, etc. and characterization (bar and pie plots).

• ClusteringAnalysis.ipynb: Source codes for motif analysis within each motif cluster: FG-motif overlaps, distribution of different binding modes, affinity comparison, etc.

• motifanalyse.py: Source codes for quantitative representation and visualization of FG-binding motifs and conservation evaluation among different ligand-binding proteins.