Benchmarking single-cell cross-omics imputation methods for surface protein expression

Datasets and scripts for executing each imputation method for "benchmarking single-cell cross-omics imputation methods for surface protein expression." The datasets are stored according to the experiments. Each experiment folder contains datasets that have undergone quality control and an intersection of genes and proteins, making them directly usable as input data for the experiments.After downloading the dataset and code, please place them according to the following directory structure: root directory │ ├── datasets │ ├── different clinical states │ ├── different protocols │ ├── ... │ └── validation without ground truth data │ ├── code │ ├── Babel │ ├── cTP-net │ ├── ... │ └── TotalVI │ └── outputs ├── different clinical states ├── different protocols ├── ... └── validation without ground truth data In code folder, we provide detailed tutorials and documentation for 12 methods. The implementation pipeline of each method is illustrated using CITE-SLN111-Gayoso: Mouse1→Mouse2 experiment in different samples scenario as example. For methods that can handle batch information within the dataset, we also provide CITE-PBMC-Li: Group1→Group2 experiment in different samples scenario as example. Note that, unless otherwise specified, the working directory for each program is root directory. The outputs folder is used to store the results of experiments.