MidaR: Small Molecule Mass Spectrometry Data Processing, Quality Control, Analysis and Re скачать в хорошем качестве

MidaR: Small Molecule Mass Spectrometry Data Processing, Quality Control, Analysis and Re

Bo Burla Analyses of small molecules, i.e., metabolites and lipids, in clinical cohorts and population studies have become valuable tools in translational research, offering new insights in metabolism, disease mechanisms, and potentially revealing prognostic, diagnostic and treatment markers of disease and health. The mostly mass spectrometry-based measurement of hundreds of small molecules in thousands of samples spanning months, however, is still challenging in all its phases. Data processing of such large-scale analyses is time-consuming and complex, and can introduce itself variability, bias, and errors. We have been implementing data processing workflows and corresponding software pipelines for the processing, management, assessment of analytical and data quality, exploratory data analysis and reporting of such large-scale analyses. These processes are mostly based on R, with in-house developed R packages, R/Shiny applications, tailored R/notebooks and our LIMS. The use of R and its novel tools helps in the rapid development, maintenance, and adaptation of software pipelines. However, the efficient and tailored handling of data from the diverse analytical methods and diverse clinical projects requires more flexibility than such implemented pipelines can provide. While these specific tasks can be managed by bioinformaticians, we also explore enabling lab members, i.e., analytical scientists and students, to process and explore their own data. We do this by conducting applied training and regular workshops on R, and by providing key functions via in-house R packages, and by supporting them accordingly. While there are challenges and limitations with this, this approach helps analytical scientists to understand their analytical results faster, interrogate them with their own questions, and allows them to identify issues in their analyses and improving analytical methods. Furthermore, the close collaboration between wet- and dry lab members contributes to a better quality of the generated data and processes in our lab. I will present examples of such workflows in our lab, including the R package ‘MidaR’ developed in our lab providing R functions and some R/Shiny tools to build reproducible data processing, quality control, exploration and report pipelines for large-scale small molecule analyses (https://github.com/SLINGhub/midar)