A Validated Analysis Pipeline for Vitreous Proteomics

Abstract

Purpose : Proteomic analysis of vitreous is capable of identifying proteins and pathways known to play critical roles in retinal disease. However, the number of samples required to generate statistically meaningful comparisons across groups remains unknown. In addition, the methods used to merge data from multiplex tandem mass tag mass spectrometry (TMT-MS) studies remains to be assessed and validated. The purpose of this study was to determine the proper algorithm to merge data across multiple 10-plexes and to identify the number of samples required to produce statistically reliable differential expression and pathway analyses.

Methods : Vitreous was derived from patients undergoing repair of epiretinal membrane or non-clearing vitreous hemorrhage secondary to proliferative diabetic retinopathy (PDR). Samples were processed according to protocols previously developed by our group. Samples were analyzed either individually (biological replicates) or as a pooled mixture of multiple samples, which was then aliquoted into smaller volumes (technical replicates). Samples were then distributed across five total 10-plexes in two TMT-MS experiments. Multiple algorithms were used to normalize the data across plexes. Validation of the normalization algorithm, differential expression, and power analyses were performed using the single normalized data matrix. This study was approved by the University of Michigan and Penn State College of Medicine institutional review boards and adhered to the tenets of the Declaration of Helsinki.

Results : The total number of unique proteins was 1,152 in experiment 1 and 1,191 in experiment 2. The average coefficient of variation (CV) across technical replicates was 4.2%. The average CV across patients was 36.9%. Power analysis revealed that 6 samples are required to achieve a power level of approximately 0.80 for proteins with log2 fold change of at least 1.194. As expected, differential expression and pathway analyses demonstrated significant activation of metabolic pathways and inhibition of neuroprotective pathways in PDR samples.

Conclusions : These data demonstrate minimal technical variability and low biological variability using TMT-MS to interrogate of human vitreous samples. In addition, methods to merge data across multiple 10-plex runs have been validated. As such, this vitreous proteomic analysis pipeline has been validated and allows for quantitative, cost-effective, and scalable analysis.

This is a 2021 ARVO Annual Meeting abstract.