PPAI: a web server for predicting protein-aptamer interactions

Abstract

Background: The interactions between proteins and aptamers are prevalent in organisms and play an important role in various life activities. Thanks to the rapid accumulation of protein-aptamer interaction data, it is necessary and feasible to construct an accurate and effective computational model to predict aptamers binding to certain interested proteins and protein-aptamer interactions, which is beneficial for understanding mechanisms of protein-aptamer interactions and improving aptamer-based therapies. Results: In this study, a novel web server named PPAI is developed to predict aptamers and protein-aptamer interactions with key sequence features of proteins/ aptamers and a machine learning framework integrated adaboost and random forest. A new method for extracting several key sequence features of both proteins and aptamers is presented, where the features for proteins are extracted from amino acid composition, pseudo-amino acid composition, grouped amino acid composition, C/T/D composition and sequence-order-coupling number, while the features for aptamers are extracted from nucleotide composition, pseudo-nucleotide composition (PseKNC) and normalized Moreau-Broto autocorrelation coefficient. On the basis of these feature sets and balanced the samples with SMOTE algorithm, we validate the performance of PPAI by the independent test set. The results demonstrate that the Area Under Curve (AUC) is 0.907 for prediction of aptamer, while the AUC reaches 0.871 for prediction of protein-aptamer interactions. Conclusion: These results indicate that PPAI can query aptamers and proteins, predict aptamers and predict protein-aptamer interactions in batch mode precisely and efficiently, which would be a novel bioinformatics tool for the research of protein-aptamer interactions. PPAI web-server is freely available at http://39.96.85.9/ PPAI.

AptaBERT: Predicting aptamer binding interactions

“…Orange points denote the true target values 0 (negative and F1-score as well as an average precision score (APP) of 96%. In comparison, the random forest-based models Li et al [45], Zhang et al [47] and PPAI [44] attained lower scores for the same metrics with Li’s model presenting the second best scores among the lineup.…”

“…Moving to the modeling of aptamer-target interactions, machine learning has been pivotal, with models emphasizing feature engineering through various methods, including pseudo-nucleotide compositions [42, 43, 44], pseudo-amino acid compositions [45, 46, 26], and the position-specific score matrix of proteins [47], supplemented by the physicochemical properties of the aptamers and proteins involved [42, 43, 44, 45, 46, 26, 47, 48, 49, 50, 51, 52]…”

“…The architectures employed range from random forests [47, 45, 44] to standard multi-layer perceptrons (MLP) [26, 53], aiming to predict binary outcomes of aptamer-protein interactions. Due to a scarcity of negative interaction data, all the aforementioned studies resorted to using randomized sequences as a control.…”

APIPred: An XGBoost-Based Method for Predicting Aptamer–Protein Interactions

“…Apart from different sequence-based methods, efficient use of feature extraction is also crucial for an accurate prediction of the aptamer–protein interaction. Li et al developed a novel feature extraction method, wherein the amino acid composition, pseudo amino acid composition, grouped amino acid composition, C/T/D composition, and sequence-order-coupling number are all employed to represent protein sequence features . Zhang et al proposed a bidirectional recurrent neural network for predicting next-generation sequencing depth from the DNA sequence, where the model utilized the local features (individual nucleotide level) as well as a small part of the global features (oligonucleotide molecule level) …”

“…Li et al developed a novel feature extraction method, wherein the amino acid composition, pseudo amino acid composition, grouped amino acid composition, C/T/D composition, and sequence-order-coupling number are all employed to represent protein sequence features. 13 Zhang et al proposed a bidirectional recurrent neural network for predicting next-generation sequencing depth from the DNA sequence, where the model utilized the local features (individual nucleotide level) as well as a small part of the global features (oligonucleotide molecule level). 7 To improve the accuracy of aptamer−protein interactions and to address these challenges, we report "APIPred", a novel model with multiple feature extraction of protein and aptamer.…”