Hypertension (HT), or high blood pressure is one of the most common and main causes in cardiovascular diseases, which is also related to a series of detrimental diseases in humans. Deficiencies in effective treatment in HT are often associated with a series of diseases including multi-infarct dementia, amputation, and renal failure. Therefore, identifying anti-hypertension peptides has the vital realistic significance. Although many bioactive peptides have been developed to reduce blood pressure, they are time-consuming and laborious. In views of the obstacles of the intrinsic methods in antihypertensive peptide (AHTP) classification, computational methods are suggested as a supplement to identify AHTPs. In this study, we develop a comprehensive feature representation algorithm based on pretrained model and convolutional neural network and apply the deep ensemble model to construct the prediction model. The new predictor is used to identify AHTPs in benchmark and independent datasets. It has been shown in the independent test set that the performance is better than the recent methods. Comparative results indicate that our model can shed some light on hypertension therapy and gains more insights of classifying AHTPs. The implements and codes can be found in https://github.com/yuanying566/AHPred-DE.

中文翻译：

---

使用预训练蛋白质嵌入识别抗高血压肽的深度集成预测器

高血压（HT）或高血压是心血管疾病中最常见和主要原因之一，它也与人类的一系列有害疾病有关。HT 有效治疗的缺陷通常与一系列疾病有关，包括多发性梗塞性痴呆、截肢和肾功能衰竭。因此，鉴定抗高血压肽具有重要的现实意义。尽管已经开发了许多生物活性肽来降低血压，但它们既费时又费力。鉴于内在方法在抗高血压肽 (AHTP) 分类中的障碍，建议使用计算方法作为识别 AHTPs 的补充。在这项研究中，我们开发了一种基于预训练模型和卷积神经网络的综合特征表示算法，并应用深度集成模型来构建预测模型。新的预测器用于识别基准和独立数据集中的 AHTP。在独立测试集中已经表明，性能优于最近的方法。比较结果表明，我们的模型可以为高血压治疗提供一些启示，并获得更多关于 AHTP 分类的见解。工具和代码可以在 比较结果表明，我们的模型可以为高血压治疗提供一些启示，并获得更多关于 AHTP 分类的见解。工具和代码可以在 比较结果表明，我们的模型可以为高血压治疗提供一些启示，并获得更多关于 AHTP 分类的见解。工具和代码可以在https://github.com/yuanying566/AHPred-DE.