基于小样本血浆蛋白质组学数据的抑郁症分类预测

Prediction of depression classification based on small sample plasma mass spectrometry data

抑郁症是一种常见的精神障碍,约27%的人在一生中会出现类似症状,早期诊断对治疗至关重要,但传统诊断方法存在主观局限性,易误诊或漏诊,因此需要一种客观的诊断方法来提高诊断率。蛋白质组学技术研究蛋白质表达水平变化,可以帮助理解疾病机制,有助于开发临床诊断工具。蛋白质组学数据通常具有特征维度高,样本量少的特点,本文提出了一种基于小样本学习的抑郁症分类预测模型,相比于传统机器学习模型,该模型对抑郁症的分类预测能力显著提升。

Depression is a common mental disorder,with about 27%of people experiencing similar symptoms during their lifetime.Early diagnosis is essential for treatment,but traditional diagnostic methods have subjective limitations that make them prone to misdiagnosis or omission,so an objective diagnostic method is needed to improve diagnosis rates.Proteomics technology studies changes in protein expression levels,which can help understand disease mechanisms and contribute to the development of clinical diagnostic tools.Proteomics data are usually characterised by high feature dimensions and small sample sizes.In this paper,we propose a classification prediction model for depression based on small sample learning,which significantly improves the classification prediction ability of depression compared to traditional machine learning models.