Skip to content

Cite seq total variational inference

CITE-seq Total Variational Inference

Data Type

scvi-tools (single-cell variational inference tools) is an analysis package for end-to-end analysis of single-cell omics data primarily developed and maintained by the Yosef Lab at UC Berkeley and the Weizmann Institute of Science.

totalVI (total Variational Inference) provides a flexible generative model of CITE-seq RNA and protein data that can subsequently be used for many common downstream tasks.

Inputs:

Combined RNA and protein data must be uploaded in the form of a MuData multimodal data object (h5mu). Batch correction will be performed on the data, and so the batch information must be included in the input data as a column in the .obs.

After the input MuData file (h5mu) has been uploaded to Cirro, the "CITE-seq Total Variational Inference (scvi-tools/totalVI)" pipeline can be selected from the pipeline catalog.

Parameters:

  • Protein Layer: Provide the key used for the protein data in the MuData object
  • RNA Layer: Provide the key used for the RNA data in the MuData object
  • Batch Key: Provide the key used for the batch information in the MuData object

Citations:

  • Adam Gayoso, Zoë Steier, Romain Lopez, Jeffrey Regier, Kristopher L Nazor, Aaron Streets, Nir Yosef (2021), Joint probabilistic modeling of single-cell multi-omic data with totalVI, Nature Methods