Sequence-Based Prediction of Plant Allergenic Proteins: Machine Learning Classification Approach

Abstract

This Article proposes a novel chemometric approach to understanding and exploring the allergenic nature of food proteins. Using machine learning methods (supervised and unsupervised), this work aims to predict the allergenicity of plant proteins. The strategy is based on scoring descriptors and testing their classification performance. Partitioning was based on support vector machines (SVM), and a k-nearest neighbor (KNN) classifier was applied. A fivefold cross-validation approach was used to validate the KNN classifier in the variable selection step as well as the final classifier. To overcome the problem of food allergies, a robust and efficient method for protein classification is needed.

pLM4Alg: Protein Language Model-Based Predictors for Allergenic Proteins and Peptides

“…This assured the universality of ESM-2 pLMs in handling our allergenic protein/peptide data sets. Moreover, since the data set contains nonstandard amino acids ([BOUZ]), the resulting ESM-2 pLMs have the ability to encode protein sequences with these residues, whereas the sequences with nonstandard residues were generally removed during model development in previous studies. ,, Furthermore, even if a given sequence has absent residues at a few positions, ESM-2 pLMs can still encode the sequence because of its inherent capacity from the masked language model strategy. , …”

“…In previous studies where global descriptors (e.g., AAC, DPC, PSSM, etc.) were used for sequence embeddings, the sequential information on proteins/peptides was lost, thereby hindering further improvement in prediction model performance. ,, For example, a tripeptide “ALL” encoded by AAC would be indistinguishable from the tripeptide “LLA”. It can be observed in the study of Sharma et al, where the pure machine learning models only achieved 84% accuracy with AAC as the descriptor .…”

“…Therefore, pLM4Alg is not recommended for prediction tasks which involve the long sequence (>900 residues) proteins. Since the traditional sequence embedding methods have significantly lower computational demand than pLMs, they are not limited by this constraint. ,,,, …”

“…Another representative technique is the machine learning (ML) approach where the classification criterion is determined by the ML model by learning from a separate training data set. ,,,− , Besides the selection of modeling method, the most challenging task in the ML approach is the representation or encoding of protein/peptide sequences into a numerical vector/matrix. Various encoding methods, including amino acid descriptors, amino acid composition (AAC), pseudoamino acid composition (PseAAC), dipeptide composition (DPC), amino acid descriptors (AAD), position-specific scoring matrix (PSSM), physicochemical descriptors, biomedical properties, k-mer dictionary-based binary representation, etc., have been widely used in predicting allergenicity and other properties/bioactivities. ,,,,,− However, these features may not always accurately represent protein sequences and simple combinations can cause high-dimensional problems as well as the feature redundancy …”

Prospects for developing allergen‐depleted food crops

“…The approach is based on rating descriptors and evaluating their classification performance. It is necessary to create a reliable and effective protein categorization system in order to overcome the issue of food allergies (Nedyalkova et al., 2023). This information can then be used to guide the selection of plant varieties and to design breeding strategies that reduce the expression of allergenic proteins or to develop targeted approaches for modifying allergenic proteins.…”