Experiences with Bioconductor-Python interoperability

Author(s): Vincent James Carey

Affiliation(s): Channing Division of Network Medicine, Harvard Medical School



Bioconductor's approach to defining and managing an R-based software/data ecosystem for computational genomic data science has led to substantial impact in academia and industry for over 20 years. The proliferation of python-based tools for machine learning and genomics has increased the importance of reliable interfaces between python and R. Posit's reticulate supports use of python modules in R; Bioconductor's basilisk strengthens the interface by encapsulating the python resources used in any given application. Efforts towards 'language agnostic' representation of genome-scale experiments are proceeding in the alabaster and gypsum frameworks in Bioconductor 3.19. In this talk I will provide an overview of the current state of interoperation through basilisk, with examples in ontology analysis and graph embedding with application to multimodal single cell data.