BrainSABER v0.99.5: transcriptomic similarity assessment toolkit and web application for the BrainSpan developmental dataset analysis

As part of their Brain Atlas project, the Allen Brain Institute has sampled the transcriptome of the developing human brain across ages ranging from 8 weeks post conception through 40 years of age for sixteen cortical and subcortical regions. The BrainSABER package is designed to provide a self-validating container for user transcriptomic data, to compare the transcriptomic similarity between the provided data and the BrainSpan dataset across developmental ages and brain regions, and to visualize the result of these comparisons. The accompanying Shiny application is designed to facilitate use of the BrainSABER toolkit and calculates the transcriptomic similarity across the 5000 genes with the highest stabilized variance across the BrainSpan dataset. Similarity metrics implemented in the BrainSABER command-line package are Euclidean and cosine distance, while the Shiny application has cosine distance, Kendall’s Tau, and Spearman’s Rho, as well as significance estimates for the correlation metrics. Results are visualized for each dataset as dynamic heatmaps in both command-line and Shiny implementations of the toolkit. Downloadable static heatmaps comparing correlation scores across datasets are also displayed in the Shiny application. The command-line BrainSABER package is published in the developmental version of Bioconductor (DOI: 10.18129/B9.bioc.BrainSABER), and the Shiny application is available in the developmental “dev” branch of the bicbioeng/BrainSABER package on GitHub as part of the ongoing developments to the published package. BrainSABER and shinyBrainSABER are designed to provide an approachable toolkit to assess and visualize similarity between user data and the Allen Brain Institute BrainSpan dataset.

Source – University of South Dakota