Re-definition of RunTSNE function from Seurat removed. This seemed to have been done to deal with errors that occurred when requesting use of more principal components than had been calculated, i.e. RunPCA() does 20 by default and the code was requesting 30 from RunTSNE(). A common specification of principal components to both routines (specified via a command arg) addressed the issue making the redefinition unnecessary.
Allows spikes to be defined via a GTF (the one automatically generated by iRAP to be supplied), such that they are removed prior to downstream analysis (which they adversely bias).
irap_tsne: General simplifying, tidying up and making code readable.
For the calling makefile irap_sc_vis.mk:
Where spikes have been specified, provide them to irap_tsne for exclusion. I have adopted this strategy (rather than removing them wholesale before the script), in case future work leverages the spike-ins.
This PR does the following.
For the irap_tsne R script:
Re-definition of RunTSNE function from Seurat removed. This seemed to have been done to deal with errors that occurred when requesting use of more principal components than had been calculated, i.e. RunPCA() does 20 by default and the code was requesting 30 from RunTSNE(). A common specification of principal components to both routines (specified via a command arg) addressed the issue making the redefinition unnecessary.
Allows spikes to be defined via a GTF (the one automatically generated by iRAP to be supplied), such that they are removed prior to downstream analysis (which they adversely bias).
irap_tsne: General simplifying, tidying up and making code readable.
For the calling makefile irap_sc_vis.mk: