digitalDLSorteR: digitalDLSorteR: an R package to deconvolute bulk RNA-Seq...

digitalDLSorteRR Documentation

digitalDLSorteR: an R package to deconvolute bulk RNA-Seq samples using single-cell RNA-Seq data and Deep Learning

Description

digitalDLSorteR is an R package that allows to deconvolute bulk RNA-Seq data using context-specific deconvolution models based on single-cell RNA-Seq data and Deep Neural Networks. These models are able to make accurate estimates of the cell composition of bulk RNA-Seq samples from the same context using the advances provided by Deep Learning and the meaningful information provided by scRNA-Seq data. See Torroja and Sanchez-Cabo (2019) (doi: 10.3389/fgene.2019.00978) for more details.

Details

The foundation of the method consists of a process that starts from single-cell RNA-Seq data and, after a few steps, a Deep Neural Network (DNN) model is trained with simulated pseudo-bulk RNA-Seq samples whose cell composition is known. These trained models are able to deconvolute any bulk RNA-Seq sample from the same biological context by determining the proportion of present cell types. The main advantage is the possibility to build deconvolution models trained with real data from certain biological environments. For example, to quantify the proportion of tumor infiltrated lymphocytes (TILs) in breast cancer, a specific model for this type of samples can be obtained by using this package. This overcomes the limitation of other methods, since stromal and immune cells may significantly change their transcriptional profiles depending on tissue and disease context.

The package can be used in two ways: deconvoluting bulk RNA-Seq samples using pre-trained models available on the digitalDLSorteRmodels R package or building your own models trained with your own single-cell RNA-Seq data. These new models may be published to make them available for other users working with similar data. So far, available models allow deconvoluting breast and colorectal cancer samples. See vignettes and https://diegommcc.github.io/digitalDLSorteR/ for more details.


digitalDLSorteR documentation built on Oct. 5, 2022, 9:05 a.m.