Hi,
I am currently exploring the application of BANKSY to multiple CODEX (PhenoCycler) datasets and would like to confirm its feasibility and best practices in this context. I have reviewed the available documentation and example workflows, but I still have a few questions specific to CODEX and TMA-style data.
My questions are:
(1) Are there any recommended preprocessing steps or parameter settings when applying BANKSY to CODEX datasets, particularly given that the data are protein-based rather than RNA-based?
(2) In our case, the data come from TMA CODEX dataset. Our current preprocessing strategy is to perform z-score normalization per marker within each TMA core, followed by Harmony to correct for batch effects between cores.
(3) When integrating multiple TMA cores using Seurat’s RunBanksy(), what would be the recommended downstream workflow?
Specifically, does the following strategy make sense?
Merge all cores into a single Seurat object + Run RunBanksy() (using TMA core as the grouping variable) + Perform dimensionality reduction on the BANKSY assay + Apply Harmony for batch correction across TMA cores
Any guidance, best practices, or example workflows for applying BANKSY to CODEX (especially TMA-based datasets) would be greatly appreciated.
Thank you very much for your time and help.
Best regards
Xianjie
Hi,
I am currently exploring the application of BANKSY to multiple CODEX (PhenoCycler) datasets and would like to confirm its feasibility and best practices in this context. I have reviewed the available documentation and example workflows, but I still have a few questions specific to CODEX and TMA-style data.
My questions are:
(1) Are there any recommended preprocessing steps or parameter settings when applying BANKSY to CODEX datasets, particularly given that the data are protein-based rather than RNA-based?
(2) In our case, the data come from TMA CODEX dataset. Our current preprocessing strategy is to perform z-score normalization per marker within each TMA core, followed by Harmony to correct for batch effects between cores.
(3) When integrating multiple TMA cores using Seurat’s RunBanksy(), what would be the recommended downstream workflow?
Specifically, does the following strategy make sense?
Merge all cores into a single Seurat object + Run RunBanksy() (using TMA core as the grouping variable) + Perform dimensionality reduction on the BANKSY assay + Apply Harmony for batch correction across TMA cores
Any guidance, best practices, or example workflows for applying BANKSY to CODEX (especially TMA-based datasets) would be greatly appreciated.
Thank you very much for your time and help.
Best regards
Xianjie