Skip to content

fitted.values is not appropriate for PCA #39

@ofarrelle

Description

@ofarrelle

df <- data.frame(lrt$fitted.values)

Our PCA plots, while useful, always look a bit funny. I have recently had more success building PCAs from variance-stabilized transformations (the vst() function in DEseq2). We should look into the most similar approach in edgeR. Currently we use the fitted.values field of the edgeR object but I think this represents just a linear transformation of the raw counts, and may still inflate the importance of high-variance genes in the PCA.

Will require a bit of research.

https://www.dnastar.com/ArrayStar_Help/index.html#!Documents/deseq2andedger.htm
https://support.bioconductor.org/p/69770/

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions