Novel computational methods for in vivo proteome dynamics estimation using heavy water metabolic labeling and LC-MS

总结 存在对可以增加体内蛋白质组覆盖率的计算方法的根本需要。 使用重水和LC-MS代谢标记的蛋白质组动力学研究。目前，只有30-40%的 所有定量的肽都用于确定蛋白质组动力学，因为其余的肽由于品质差而被过滤， 实验数据与其理论拟合之间的拟合度（皮尔逊相关性，残差平方和）。很长的- 长期目标是开发方法，用于推断蛋白质周转变化对 疾病的发展。本申请的目的是开发计算方法， 蛋白质周转使用丰度只有两个质量同位素，估计可交换的数量 来自三种质量同位素异构体肽中的氢，并使用色谱比对来定量标记 在MS 2中未对洗脱曲线取样的肽中掺入。目前的估计方法 蛋白质周转使用肽的单一同位素峰的相对丰度（RA）消耗的时间过程， 如由MS 1中肽的完整同位素分布的标准化确定的。因此，只有肽 在MS 2中鉴定的用于定量标记掺入。RA的确定需要准确 肽的所有同位素异构体（≤ 6）的定量。在复杂的样品中，至少一种 通过共洗脱物质的同位素异构体的含量高。氘掺入肽的模型依赖于 可交换氢的数量，NEH。NEH值已被准确地确定只有一只小鼠。 在初步数据的基础上，将努力实现三个具体目标，以解决方法学问题： 开发、测试和验证生物信息学解决方案，以使用两种质量 同位素异构体; 2）开发，测试和验证计算方法，以估计可交换的 氢从三个质量同位素异构体;和3）开发生物信息学解决方案，以解决缺失的数据 代谢标记存在的问题。我们导出了两个新的方程， 代谢标记中三种质量同位素的丰度。目标1的基本原理是， 两种质量同位素异构体的原始丰度可用于从原始丰度估计标记物定量 只有两种质量同位素的丰度。目标2的基本原理是，三个质量的两个方程 同位素异构体可用于估计NEH值。Aim 3使用了 色谱图，以获得时间扭曲函数。基本原理是互信息是更好的- 用于估计肽在不同时间点的曲线之间的非线性关系的合适标准 代谢标记。 目标1和3将提供生物信息学解决方案，增加重水标记中的蛋白质组覆盖率， LC-MS实验。目标2的实施将使该技术应用于具有以下功能的系统成为可能： 未知氨基酸NEH值。

SUMMARY There is a fundamental need for computational methods that can increase proteome coverage for in vivo studies of proteome dynamics using metabolic labeling with heavy water and LC-MS. Currently, only 30-40% of all quantified peptides are utilized to determine proteome dynamics, as the rest are filtered due to poor goodness- of-fit (Pearson correlation, residual sum of squares) between experimental data and its theoretical fit. The long- term goal is to develop methods for inferring the causative effects of protein turnover changes on the development of diseases. The objectives of this application are to develop computational methods to estimate protein turnover using abundances of only two mass isotopomers, estimate the number of exchangeable hydrogens in a peptide from three mass isotopomers, and use chromatogram alignment to quantify label incorporation into peptides whose elution profiles have not been sampled in MS2. Current methods for estimation of protein turnover use time-course of relative abundance (RA) depletion of the monoisotopic peak of a peptide, as determined from a normalization of the complete isotope profile of the peptide in MS1. Thus, only peptides identified in MS2 are used in quantification of label incorporation. Determination of the RA requires accurate quantification of all isotopomers (≤ six) of a peptide. In complex samples, contamination of at least one of the isotopomers by a co-eluting species is high. The model of deuterium incorporation into a peptide is dependent on the number of exchangeable hydrogens, NEH. NEH values have been accurately determined only for a mouse. Based on preliminary data, three specific aims will be pursued to resolve the methodological issues: 1) Develop, test, and validate bioinformatics solutions to determine degradation rate constant using two mass isotopomers; 2) Develop, test, and validate computational methods to estimate the number of exchangeable hydrogens from three mass isotopomers; and 3) Develop bioinformatics solutions to address the missing data problem in the presence of metabolic labeling. We derived two new equations relating the time-course of raw abundances of three mass isotopomers in metabolic labeling. The rationale for Aim 1 is that the equation relating the raw abundances of two mass isotopomers can be used to estimate label quantification from the raw abundances of only two mass isotopomers. The rationale for Aim 2 is that the two equations for three mass isotopomers can be used to estimate the NEH values. Aim 3 uses mutual information between the chromatographic profiles to obtain a time-warping function. The rationale is that mutual information is a better- suited criterium for estimation of the non-linear relationship between profiles of a peptide at different timepoints of metabolic labeling. Aims 1 and 3 will provide bioinformatics solutions that increase proteome coverage in heavy water labeling and LC-MS experiments. The implementation of Aim 2 will make it possible to apply the technology to systems with unknown amino acid NEH values.