HyperCys: A Structure- and Sequence-Based Predictor of Hyper-Reactive Druggable Cysteines

Contact: * mingjie.gao@pharmazie.uni-freiburg.de * stefan.guenther@pharmazie.uni-freiburg.de

Keyword: machine learning structure- and sequence-based druggable cysteine reactivity prediction

Features Collection

• Feature Collection from 3D structures (34 features)
  • Pocket profile (PP)
    • Fpocket
      • Fpocket(https://bmcbioinformatics.biomedcentral.com/articles/10.1186/1471-2105-10-168) was used to determine whether or not amino acids are located in detectable binding pockets, and only those cysteines located in pockets were retained. Druggability, hydrophobicity, and polarity were collected.
    • Depth
      • Depth (https://academic.oup.com/nar/article/41/W1/W314/1111943?login=false) was used to calculate the depth of cysteine burial.
  • SASA-PKA profile (SPP)
    • The pKa and SASA values for each cysteine and its surrounding amino acids within 4, 6, 8, and 10 Å of cysteine were calculated using PROPKA 3 (https://pubmed.ncbi.nlm.nih.gov/26596171/) and freeSASA (https://freesasa.github.io/), respectively.
  • Amino Acid Composition (AAC)
    • Total number of surrounding amino acid residues within 4, 6, 8, and 10 Å of cysteine was calculated, and also the respective number of polar and hydrophobic amino acids according to the hydrophobic/polar (H/P) classification by Kamtekar et al. (https://pubmed.ncbi.nlm.nih.gov/8259512/).
    • Canonical H/P patterning: ADEGHKNPQRST-CFILMVWY
• Feature Collection from sequences (46 features)
  • HyperCys utilizes DisPredict2 to extract sequence features.
  • DisPredict_v2.0 (http://cs.uno.edu/~tamjid/Software.html).
  • Sumaiya Iqbal and Md Tamjidul Hoque, 2016 Contact:
    • thoque@uno.edu
    • siqbal1@my.uno.edu (anni.are@gmail.com)

1. Column 10 - 29: 20 PSSMs

    			- generate with PSI-BLAST and normalize using 9.0 -	
   

2. Column 30 - 32: 3 secondary structure (helix, strand and coil) probabilities

    			- generate using SPINE X -
   

3. Column 33: 1 accessible surface area

    			- generate using SPINE X -
   

4. Column 36 - 56: 1 monogram and 20 bigrams

    		- generate using source code provided with DisPredict supplementary materials
   
    		- available at SuppMaterial_DisPredict/Software/Source Codes
   
    		- guidelines to execute it is described in "Steps to prediction" below
   
    		- normalize the values with exponent (6.0)
   

5. Column 57 1 Position Specific Estimated Energy (PSEE)

• Totally collected:
  • 20 PSSMs
  • 1 monogram
  • 20 bigrams
  • 1 PSEE
  • 1 Helix probability
  • 1 Beta-Strand probability
  • 1 Coil probability
  • 1 ASA

Folders

• Data Folder
  • Storing bechmark data for training the machine learning models.
• Code Folder

  • code_for_sequence_based_features_collection

  • code_for_3d_structure_features_collection

  • code_for_prediction

• Model Folder

  • HyperCys.model

  • sequence_based.model

  • structure_based.model

Cite Us