AllergenAI: a deep learning model predicting allergenicity based on protein sequence


bullet pointAllergenAI overview
  • Innovations in protein engineering can help redesign allergenic proteins to reduce adverse reactions in sensitive individuals. To accomplish this aim, a better knowledge of the molecular properties of allergenic proteins and the molecular features that make a protein allergenic is needed. We present a novel AI-based tool, AllergenAI, to quantify the allergenic potential of a given protein. Our approach is solely based on protein sequences, differentiating it from previous tools that use some knowledge of the allergens' physicochemical and other properties in addition to sequence homology. We used the collected data on protein sequences of allergenic proteins as archived in the three well-established databases, SDAP 2.0, COMPARE, and AlgPred 2, to train a convolutional neural network and assessed its prediction performance by cross-validation. We then used Allergen AI to find novel potential proteins of the cupin family in date palm, spinach, maize, and red clover plants with a high allergenicity score that might have an adverse allergenic effect on sensitive individuals. By analyzing the feature importance scores (FIS) of vicilins, we identified a proline-alanine-rich (P-A) motif in the top 50% of FIS regions that overlapped with known IgE epitope regions of vicilin allergens. Furthermore, using~ 1600 allergen structures in our SDAP database, we showed the potential to incorporate 3D information in a CNN model. Future, incorporating 3D information in training data should enhance the accuracy. AllergenAI is a novel foundation for identifying the critical features that distinguish allergenic proteins.

    • bullet pointTraining and validation data
    Training data
    - one-hot encode protein matrix (allergens, positive)
    - protein information index in the one-hot matrix (allergens, positive)
    - one-hot encode protein matrix (non-allergens, negative)
    - protein information index in the one-hot matrix (non-allergens, negative)

    Cupin proteins in SDAP2.0
    - one-hot encode protein matrix (cupin protiens in SDAP)
    - protein information index in the one-hot matrix (cupin protiens in SDAP)

    Non-allergen proteins in the cupin pfam
    - one-hot encode protein matrix (non-allergenic cupin)
    - protein information index in the one-hot matrix (non-allergenic cupin)

    Modeling with SDAP allergens (with protein sequence and protein 3D structure information)
    - one-hot encode protein matrix (allergens, positive)
    - protein information index in the one-hot matrix (allergens, positive)
    - one-hot encode protein matrix (non-allergens, negative)
    - protein information index in the one-hot matrix (non-allergens, negative)

    Modeling with SDAP allergens (with protein sequence, but without protein 3D structure information)
    - one-hot encode protein matrix (allergens, positive)
    - protein information index in the one-hot matrix (allergens, positive)
    - one-hot encode protein matrix (non-allergens, negative)
    - protein information index in the one-hot matrix (non-allergens, negative)

    bullet pointAvailable Models and Codes for the AllergenAI
  • AllergenAI codes
    - AllergenAI model with full training data
    - a model with SDAP allergen protein sequence information
    - a model with SDAP allergen protein sequence and 3D structure information

  • Run Pre-processing of the input protein sequence
    - Pre-process: make one-hot encode protein matrix
    Command: python AllergenAI_preprocess.py input.fata
    Example command: python AllergenAI_preprocess.py Cupin.fasta
    Example fasta file: Cupin.fasta

  • Run AllergenAI
    - Predict the allergenicity of your protein by running AllergenAI model
    Command:pyton Run_AllergenAI.py input.txt
    Example command: pyton Run_AllergenAI.py Cupin.txt
    Example concatenated one-hot encode matrix of your proteins (made in the pre-processing step): Cupin.txt
    Output: P of non-allergen, P of allergen

  • Requirements
    Software and algorithms to train and run AllergenAI
    - python3
    - packages: tensorflow, keras2.11, numpy and pandas
    # install python
conda update conda
conda create -n allergenai python=3
conda activate allergenai

# install requirements
conda install numpy
conda install pandas
conda install tensorflow
pip install -upgrade tensorflow
conda install keras2.11


    • bullet point About us
  • Pora Kim, MS, Ph.D.
  • Email: Pora.Kim@uth.tmc.edu
  • Mailing address:
    •   Center for Computational Systems Medicine
      Department of Bioinformatics and Systems Medicine
      McWilliams School of Biomedical Informatics
      The University of Texas Health Science Center at Houston
      7000 Fannin Street, Houston, TX 77030

  • Werner Braun, Ph.D.
  • Email: webraun@utmb.edu
  • Mailing address:
    •   Department of Biochemistry and Molecular Biology
      Sealy Center for Structural Biology and Molecular Biophysics
      University of Texas Medical Branch
      301 University Blvd, Galveston, TX 77555
    bullet point Related citations

        - Ivanciuc O, Schein H C, Braun W. SDAP: database and computational tools for allergenic proteins. Nucleic Acids Res. 2003 Jan 1;31(1):359-62. doi: 10.1093/nar/gkg010.
        - Negi S S, Schein H C, Braun W. The updated Structural Database of Allergenic Proteins (SDAP 2.0) provides 3D models for allergens and incorporated bioinformatics tools. J Allergy Clin Immunol Glob. 2023 Aug 11;2(4):100162. doi: 10.1016/j.jacig.2023.100162.