Interface to "multomics" model described in paper

[mmyers1]

I have multimodal 10x data (both RNA-seq and ATAC-seq measured from the same cells) and would like to use the "multomics" model (Fig. 2C in the PLoS Comp Bio paper) in which each cell is assigned to the same cluster in both modalities. However, the vignette for fitting a CONGAS+ model returns different cluster sizes between the two modalities, and I can't find a parameter to e.g. init() or fit_congas() to specify the multiomics model or that clusters should be consistent between modalities. Is there an interface available to the multiomics model?

[Militeee]

Correct me if I am wrong @lucreziaPatruno but setting the parameter same_mixing = TRUE in the fit_congas function should be enough to force the model to use the same assignments

[lucreziaPatruno]

Hi @mmyers1, in addition to setting same_mixing = TRUE in the fit_congas function, you will need to re-run the auto_config_run function setting the parameter multiome = TRUE. Please update your installation of the Rcongas package, as I just submitted a fix for a minor bug related to the multiome model. BW Lucrezia

[mmyers1]

Thanks! Using those instructions I was able to cluster the cells in the demo data consistently between the two modalities.