An interactive R-based open-source solution for qualitative and quantitative lipidomics of bovine milk

Abstract: Commercially available software for data analysis is expensive and typically needs to be constantly updated at substantial cost. But we now live in an era when open-source solutions for data science have become routinely available for free use. The programming language 'R' is an example of a new tool that was first distributed in the year 2000, and which has undergone a dramatic rate of development of targeted sub-routines called 'packages'. Extremely popular online software platforms such as XCMS Online and MetaboAnalyst that are used for metabolomics now have standalone, offline versions available that can reproduce the results obtained online, but without some of the drawbacks that the online platforms had, such as file size limitations, and limiting the number of projects that can be uploaded. We report here a new set of program scripts that allow quantitative analysis that is not allowed using the popular online platforms. We provide solutions to quantify 37 triacylglycerols (TAGs) using linear calibration, polynomial fit calibration, power fit calibration, response factor quantification, and bracketed linear extrapolation and double-bracketed linear extrapolation, with the latter proving most effective. Differences between the milk from the Jersey cow and the Holstein cow on the control diet were evident, such as 58.02 ± 11.42 µg/mL of C30 TAGs versus 27.62 ± 2.32 µg/mL, respectively. The Jersey cow that was fed corn increased C30 to 72.74 ± 21.01 µg/mL, while the Holstein cow fed alfalfa increased C30 to 39.13 ± 4.79 µg/mL. Additional data analysis is ongoing. We also demonstrate what has now become conventional lipidomic analysis using online databases, such as XCMS online and MS-DIAL. We report scripts that allow data processing results to be converted directly to web pages with interactive graphics, so the results can be verified or reproduced by anyone.