Datasets

Tutorial datasets

We provided the processed h5ad files used in all tutorials:

• fewshot_sample_adt.h5ad: fewshot_sample_adt

• fewshot_sample_rna.h5ad: fewshot_sample_rna

• test_sample_adt.h5ad: test_sample_adt

• test_sample_rna.h5ad: test_sample_rna

• train_sample_adt.h5ad: train_sample_adt

• train_sample_rna.h5ad: train_sample_rna

• valid_sample_adt.h5ad: valid_sample_adt

• valid_sample_rna.h5ad: valid_sample_rna

Public datasets

We provide links to the original public datasets used in our paper:

• BM (CITE-seq): GSE128639

• BMMC (CITE-seq): 194122

• CBMC (CITE-seq): GSE10086

• PBMC (REAP-seq): GSE100501

• Perturb (ECCITE-seq): GSE15306

• Heart (CITE-seq): GSE218392

• Spatial (10X Visium): 10X Visium gene-protein expression library of human tonsil

Pretrained models

We provide pretrained scLinguist models:

• decoder.ckpt: Deocder

• encoder.ckpt: Encoder

• finetune.ckpt: Finetuned