In this issue, vol. 27, issue 6, June 2023, 4 papers are published related to the to the topic Bioinformatics Informatics. Please click here to view them, with link in IEEE XPLORE.
An Effective Model for Predicting Phage-host Interactions via Graph Embedding Representation Learning with Multi-head Attention Mechanism.
The phage therapy requires to predict accurately potential phage-host associations. Some computational methods are proposed to predicte associations between phages and hosts. However, these methods ignored that the association network usually sparse. Which may influence the accuracy of phagehost prediction. In the treatment of bacterial infectious diseases, overuse of antibiotics may promote emergence of bacterial resistance and dysbiosis of beneficial bacteria, where these beneficial bacteria maintain the normal life activities of the human body. In contrast, phage therapy invades and lyses specific pathogenic bacteria without the effect on beneficial bacteria. Therefore, phage therapy is now replacing antibiotic therapy to treat bacterial infectious diseases. The potential phage-host associations play an essential role in phage therapy. Considering that a phage only infects specific bacteria (i.e., hosts), the phage-host associations network is generally sparse and non-connected. However, existing phage-host association prediction methods ignored the non-connection and sparsity of the graph caused by differences in phage-specificity. We propose a novel graph embedding representation learning based on multi-head attention mechanism for predicting phage-host interactions (GERMAN-PHI). The multi-head attention mechanism can be used to learn the embedding representations from different specificity principles with multiple perspectives, which ensures the model can predict the associations between phages and bacteria from different connected fractions. Firstly, we use GERMAN-PHI to learn the embedding feature representations of phages and bacteria. Then the embedded features are as the inputs of the neural induction matrix completion to reconstruct the associations matrix. Compared with the state-of-the-art methods, GERMAN-PHI showed a more excellent performance in terms of receiver operating characteristic curve area, suggesting that GERMAN-PHI outperforms existing methods in phage-host association prediction.
Zhao, Weizhong; Wang, Yue; Sun, Han; Wang, Haodong; Li, Dandan; Jiang, Xingpeng; Shen, Xianjun.
Predicting CircRNA-Disease associations via feature convolution learning with heterogeneous graph attention network.
Peng, Li; Yang, Cheng; Chen, Yifan; Liu, Wei.
Exploratory Analysis of the Gene Expression Matrix based on Dual Conditional Dimensionality Reduction.
Dνaz, Ignacio; Enguita, Josι M.; Cuadrado, Abel A.; Garcνa, Diego; Gonzαlez, Ana; Valdιs, Nuria; Chiara, Marνa D.
Deep Domain Adaptation Enhances Amplification Curve Analysis for Single-Channel Multiplexing in Real-Time PCR.
Xu, Ke; Mao, Ye; Miglietta, Luca; Kreitmann, Louis; Moser, Nicolas; Georgiou, Pantelis; Holmes, Alison; Rodriguez-Manzano, Jesus.