NEPTUNE

NEPTUNE是基于支持向量机的肿瘤归巢肽 (THPs) 分类器，该分类器来自最佳基线模型生成的概率信息。Thp是短肽 (即3-30个氨基酸)。它用于肿瘤诊断和治疗应用，如肿瘤部位药物递送。NEPTUNE是堆叠集成学习方法，能够确保THPs预测准确性，同时与现有方法 (即THpred，SCMTHP和MIMML) 进行比较。它的两层预测模型，在第一层中，它产生来自不同集合模型 (称为基线模型) 的概率，以及那些经过具有堆叠策略的元建模器以提高预测精度的概率。NEPTUNE is a support vector machine-based classifier for tumor-homing peptides (THPs) that utilizes probability outputs from optimal baseline models. THPs are short peptides comprising 3 to 30 amino acids, which have applications in tumor diagnosis and therapy such as targeted drug delivery to tumor sites. As a stacked ensemble learning framework, NEPTUNE ensures high prediction accuracy for THPs and is benchmarked against existing state-of-the-art methods including THpred, SCMTHP, and MIMML. Its two-layer predictive workflow first generates probability scores from diverse base models (referred to as baseline models), which are then integrated by a meta-modeler adopting a stacking strategy to further improve prediction accuracy.

搜集汇总

背景与挑战

背景概述

NEPTUNE是一个基于支持向量机的肿瘤归巢肽分类数据集，采用两层堆叠集成学习方法提升预测精度，并与现有THpred等方法进行比较。该数据集由Protiens & Peptides Mining Lab于2022年发布，主要用于肿瘤诊断和治疗应用。