bandeiralab/Pep2Prob

Pep2Prob数据集是一个全面的肽段特异性碎片离子概率预测数据集，用于串联质谱（MS/MS）基础上的蛋白质组学研究。该数据集通过基于肽序列上下文预测碎片化概率，解决了传统全局统计方法的局限性。数据集包含610,117个独特肽前体的碎片离子概率统计数据，来源于超过1.83亿个高分辨率HCD光谱。它具有多样的前体表示，不同的长度和电荷状态，高质量注释，并采用了一种新颖的训练-测试分割方案，以最小化训练集和测试集之间的结构相似性。

Pep2Prob is a comprehensive dataset designed for predicting peptide-specific fragment ion probability in tandem mass spectrometry (MS/MS) based proteomics studies. It addresses the limitations of conventional global statistical approaches by enabling the development of models that can predict fragmentation probabilities based on peptide sequence context. The dataset includes fragment ion probability statistics for 610,117 unique peptide precursors derived from over 183 million high-resolution HCD spectra, with diverse representation of precursors in terms of length and charge state, high-quality annotations, and a novel train-test split scheme to minimize structural similarity between the training and testing sets.