Giters

satijalab / seurat

R toolkit for single cell genomics

Home Page:http://www.satijalab.org/seurat

Repository from Github https://github.comsatijalab/seuratRepository from Github https://github.comsatijalab/seurat

Normalisation / SCTransform for SVG spatial analysis

cathalgking opened this issue · comments

commented

I have a 10x visium dataset within my seu object and I want to run Seurat::FindSpatiallyVariableFeatures on this data. It seems that this function first requires the data to be normalised with Seurat::SCTransform prior to calculation. Not doing so results in this error message:

Error in `object[[assay]]`:
! ‘SCT’ not found in this Seurat object

Is it possible to run FindSpatiallyVariableFeatures on data that is not normalised? Or on data that is normalised with an alternative normalisation method such as Seurat::NormalizeData for example?

commented

Hi! We recommend normalizing the data before going through the FindSpatiallyVariableFeatures workflow - you can read more about some of the rationale for suggesting SCTransform over other methods such as LogNormalize in the vignette here (https://satijalab.org/seurat/articles/spatial_vignette). Can you provide some more context on what the structure of your object looks like or how you are hoping to analyze the data without normalization?

commented

Hi @mccormickceci
It is more to be able to run the function Seurat::FindSpatiallyVariableFeatures with an alternative method of normalisation other than SCT. Is this possible?