CancerSTFormer datasets, pretrained models, and finetuned models

Datasets:

CancerSTFormer_TNBC_Normal_filtered.dataset:

This is the TNBC Visium ST collection, tokenized according to the specifications of the CancerSTFormer-50um Local Model.

CancerSTFormer_TNBC_neighbor.dataset:

This is the TNBC Visium ST collection, tokenized according to the specifications of the CancerSTFormer-250um Extended Model.

Models:

pretrained_CancerSTFormer_250um:

Pretrained model for the 250um-Extended Model.

pretrained_CancerSTFormer_50um:

Pretrained model for the 50um-Local Model.

Pickle Files for Gene Dictionaries, etc:

250um Extended Model: Please get the pickle files from https://github.com/bernard2012/CancerST/tree/main/example/extended.perturb. You need gene_median_dict.pickle, gene_name_id_dictionary.pickle, and new_token_dictionary.pickle.

50um Local Model: Please get the pickle files from: https://github.com/bernard2012/CancerST/tree/main/example/local.perturb. You need gene_median_dict.pickle, gene_name_id_dictionary.pickle, and new_token_dictionary.pickle.

More information about how to use these datasets and models for various downstream analyses and transfer learning applications can be found at https://github.com/bernard2012/CancerST/.

Website about usage and supplementary materials associated with the paper can be found at: https://qianzhulab.github.io/suppl/CancerST/.