一种基于加权非负最小二乘的蛋白质定量方法

A protein quantification method based on weighted non-negative least squares

目的提出一种综合利用肽段信息的蛋白质定量方法WeQuant(weighted peptide-based protein quantification),以提高蛋白质定量的通量与准确度,特别是低丰度蛋白质。方法基于肽段和蛋白质的关系,按照肽段的丰度与来源对所有肽段进行加权,并利用肽段和蛋白质的等量关系建立加权非负最小二乘模型,从而得到蛋白质的相对丰度。结果与传统蛋白质定量方法相比,WeQuant在实验数据集上显著增加了有效定量的蛋白质数量,并在不同丰度范围均达到了更高的定量准确度。此外,WeQuant能够有效定量未被其他方法报告的低丰度蛋白质。结论本文提出的基于加权非负最小二乘的模型能够克服对高丰度肽段和唯一肽段的依赖,实现对不同丰度范围的蛋白质进行准确定量。

Objective A protein quantification method named WeQuant(weighted peptide-based protein quantification),which comprehensively utilizes peptide information, is proposed to improve the throughput and accuracy of protein quantification, especially for low-abundance proteins.Methods Based on the relationship between peptides and proteins, all peptides are weighted according to their abundances and sources, and a weighted non-negative least squares model is established using the equal relationship between peptides and proteins to obtain the relative abundance of proteins.Results Compared with traditional protein quantification methods, WeQuant significantly increases the number of effective quantitative proteins in the experimental data set, and achieves higher quantitative accuracy in different abundance ranges.In addition, WeQuant effectively quantified low-abundance proteins not reported by other methods.Conclusions The model based on weighted non-negative least squares proposed in this paper can overcome the dependence on high-abundance peptides and unique peptides, and achieve accurate quantification of proteins with different abundance ranges.