RESUMEN
Data-independent acquisition (DIA) has improved the identification and quantitation coverage of peptides and proteins in liquid chromatography-tandem mass spectrometry-based proteomics. However, different DIA data-processing tools can produce very different identification and quantitation results for the same data set. Currently, benchmarking studies of DIA tools are predominantly focused on comparing the identification results, while the quantitative accuracy of DIA measurements is acknowledged to be important but insufficiently investigated, and the absence of suitable metrics for comparing quantitative accuracy is one of the reasons. A new metric is proposed for the evaluation of quantitative accuracy to avoid the influence of differences in false discovery rate control stringency. The part of the quantitation results with high reliability was acquired from each DIA tool first, and the quantitative accuracy was evaluated by comparing quantification error rates at the same number of accurate ratios. From the results of four benchmark data sets, the proposed metric was shown to be more sensitive to discriminating the quantitative performance of DIA tools. Moreover, the DIA tools with advantages in quantitative accuracy were consistently revealed by this metric. The proposed metric can also help researchers in optimizing algorithms of the same DIA tool and sample preprocessing methods to enhance quantitative accuracy.
Asunto(s)
Proteómica , Espectrometría de Masas en Tándem , Proteómica/métodos , Proteómica/normas , Proteómica/estadística & datos numéricos , Espectrometría de Masas en Tándem/métodos , Espectrometría de Masas en Tándem/normas , Cromatografía Liquida/métodos , Cromatografía Liquida/normas , Algoritmos , Reproducibilidad de los Resultados , Humanos , Benchmarking , Péptidos/análisis , Programas Informáticos , Cromatografía Líquida con Espectrometría de MasasRESUMEN
RATIONALE: In clinical diagnosis of liver injury, which is an important health concern, serum aminotransferase assays have been the go-to method used worldwide. However, the measurement of serum enzyme activity has limitations, including inadequate disease specificity and enzyme specificity. METHODS: With the high selectivity and specificity provided by nano liquid chromatography-tandem mass spectrometry (LC/MS/MS), this work describes a method for the simultaneous determination of six proteins in liver that can be potentially used as biomarkers for liver injury: glutamic-pyruvic transaminase 1 (GPT1), glutamic oxaloacetic transaminase 1 (GOT1), methionine adenosyl transferase 1A (MAT1A), glutathione peroxidase 1 (GPX1), cytokeratin 18 (KRT18) and apolipoprotein E (APOE). RESULTS: In validation, the method was shown to have good selectivity and sensitivity (limits of detection at pg/mL level). The analytical method revealed that, compared with normal mice, in carbon tetrachloride-induced acute liver injury mice, liver MAT1A and GPX1 were significantly lower (p < 0.01 and p < 0.05, respectively), KRT18 was significantly higher (p < 0.05) and APOE and GPT1 were marginally significantly lower (p between 0.05 and 0.1). This is the first work reporting the absolute contents of GPT1, GOT1, MAT1A, GPX1 and KRT18 proteins based on LC/MS. CONCLUSIONS: The proposed method provides a basis for establishing more specific diagnostic indicators of liver injury.
Asunto(s)
Hígado , Espectrometría de Masas en Tándem , Animales , Ratones , Espectrometría de Masas en Tándem/métodos , Cromatografía Liquida/métodos , Hígado/metabolismo , Apolipoproteínas E/metabolismoRESUMEN
Proteomics provides molecular bases of biology and disease, and liquid chromatography-tandem mass spectrometry (LC-MS/MS) is a platform widely used for bottom-up proteomics. Data-independent acquisition (DIA) improves the run-to-run reproducibility of LC-MS/MS in proteomics research. However, the existing DIA data processing tools sometimes produce large deviations from true values for the peptides and proteins in quantification. Peak-picking error and incorrect ion selection are the two main causes of the deviations. We present a cross-run ion selection and peak-picking (CRISP) tool that utilizes the important advantage of run-to-run consistency of DIA and simultaneously examines the DIA data from the whole set of runs to filter out the interfering signals, instead of only looking at a single run at a time. Eight datasets acquired by mass spectrometers from different vendors with different types of mass analyzers were used to benchmark our CRISP-DIA against other currently available DIA tools. In the benchmark datasets, for analytes with large content variation among samples, CRISP-DIA generally resulted in 20 to 50% relative decrease in error rates compared to other DIA tools, at both the peptide precursor level and the protein level. CRISP-DIA detected differentially expressed proteins more efficiently, with 3.3 to 90.3% increases in the numbers of true positives and 12.3 to 35.3% decreases in the false positive rates, in some cases. In the real biological datasets, CRISP-DIA showed better consistencies of the quantification results. The advantages of assimilating DIA data in multiple runs for quantitative proteomics were demonstrated, which can significantly improve the quantification accuracy.