imput of Bulk RNA-seq have negative after romoving batch effect

[Lily159753]

Dear Tinyi Chu:

Hello!

My input of the Bulk RNA-seq is the combination of 3 batches, so I perform the cambat from SVM R packages to romve the batch effect. However, despite the original data are all counts, there are negative numbers after removing batches.

To better perform the BayesPrism, can you give me some advice on how to overcome or choose other methods for my bulk RNA-seq preprocessing?

Thank you so much!

[tinyi]

Thank you for your interest in our methods. I would recommend direct run BayesPrism without performing batch effect correction on the bulk, as BayesPrism is highly robust to linear batch effects (you may refer to our manuscript for details on this). If there are significant non-linear batch effects, such as biological variation in cell type-specific gene expression, you may also deconvolve each batch separately.

…

On Wed, Jan 10, 2024 at 6:34 AM Lily159753 ***@***.***> wrote: Dear Tinyi Chu: Hello! My input of the Bulk RNA-seq is the combination of 3 batches, so I perform the cambat from SVM R packages to romve the batch effect. However, despite the original data are all counts, there are negative numbers after removing batches. To better perform the BayesPrism, can you give me some advice on how to overcome or choose other methods for my bulk RNA-seq preprocessing? Thank you so much! — Reply to this email directly, view it on GitHub <#75>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB4NHSYUS6C323PPRQISWPDYNZ4CXAVCNFSM6AAAAABBUUGUH6VHI2DSMVQWIX3LMV43ASLTON2WKOZSGA3TIMJXGU3TSMY> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>