MS-Helios: a Circos wrapper to visualize multi-omic datasets

Harald Marx,Joshua J Coon

doi:10.1186/s12859-018-2564-9

Abstract

BackgroundAdvances in high-resolution mass spectrometry facilitate the identification of hundreds of metabolites, thousands of proteins and their post-translational modifications. This remarkable progress poses a challenge to data analysis and visualization, requiring methods to reduce dimensionality and represent the data in a compact way. To provide a more holistic view, we recently introduced circular proteome maps (CPMs). However, the CPM construction requires prior data transformation and extensive knowledge of the Perl-based tool, Circos.ResultsWe present MS-Helios, an easy to use command line tool with multiple built-in data processing functions, allowing non-expert users to construct CPMs or in general terms circular plots with a non-genomic basis. MS-Helios automatically generates data and configuration files to create high quality and publishable circular plots with Circos. We showcase the software on large-scale multi-omic datasets to visualize global trends and/or to contextualize specific features.ConclusionsMS-Helios provides the means to easily map and visualize multi-omic data in a comprehensive way. The software, datasets, source code, and tutorial are available at https://sourceforge.net/projects/ms-helios/.

Highlights

Advances in high-resolution mass spectrometry facilitate the identification of hundreds of metabolites, thousands of proteins and their post-translational modifications
Cost-effective, and comprehensive data acquisition methods, systems biology is undergoing a transition from single-omic to multi-omic data analysis [1]
To provide a holistic and integrated view, we recently introduced circular proteome maps (CPMs), visualizing sample features in a circular plot in a proteome-centric way [4]

Summary

Results

MS-Helios workflow MS-Helios builds circular plots with a non-genomic basis from datasets in delimited text file format, where rows represent features and columns samples (Fig. 1). The default configuration of MS-Helios and Circos enables users to produce high quality and publishable figures, requiring minimal input from the user to build the data and config files for Circos. Each ideogram cluster illustrates proteins by organ occurrence, e.g., the first cluster (Fig. 2a, blue bar) contains 1872 proteins present in five organs known as core proteome. To explore protein expression in the core and specific proteome (Fig. 2a, green bars), we utilize the built-in scaling normalization method (Fig. 2a, black histogram). By mapping the transcript data to the proteomes (red histogram), we are able to illustrate similar trends in the core but the opposite for specific clusters. Each individual cluster illustrates that high abundant proteins correlate with high abundant transcripts, but this trend is not generalizable for the complete cluster (Fig. 2b)

Background

Conclusion