Recently, the ability to profile gene expression at the single-cell level has been expanded to include the measure of both RNA and protein levels from the same cells (Stoeckius et al., 2017). While RNA-based data has broadly proven sufficient for cell type detection and classification, its usage for determining drug target is still limiting: we still often rely on the assumption that RNA expression levels are representative of protein expression levels. In this presentation I will summarize the results of a six-months internship at Roche that aimed to bridge the gap between the RNA and the protein worlds.
In the first part of the presentation, I will briefly present Besca (Maedler et al., 2020), a single-cell analysis workflow based on scanpy that allows to process raw single-cell RNA-seq data to annotated clusters. Besca was recently extended to CITE-seq analysis to provide a simple multimodal analysis workflow for this type of data.
In the second part of the presentation, I will present a novel neural network, cTP-net (Zhou et al., 2020), for the prediction of protein expression from raw single-cell RNA-seq data. cTP-net is able to learn the relation between RNA and protein levels and provides accurate predictions in different settings. I will present the network architecture, the enhancements that we brought to it and example applications.