INSISTC require the following package with python 3.6 or higher:
  numpy=1.19.2
  pandas=1.1.5
  scikit-learn=0.24.2
  loompy=3.0.6
  scipy=1.5.2
  networkx=2.5.1
  pyscenic=0.11.2

We provided a conda requirements file, which can be used to create conda environment: "conda create --name <env_name> --file <requirements_file_name>"

Memory requirement: 32GB or above(may consumer more memory depends on gene regulatory network size and the sizes of expression matrix)


INSISTC can run with 3 different mode: 
1. The default running mode using gene-walk-based vocabulary(--running_mode 0). The input for this mode include a file specify an a regulatory network, a file specify random walk anchor. Loom file contain scRNA-Seq expression matrix(The row are "Gene" and the column are "CellID"). 
2. Walk-only running mode(--running_mode 1), use only walk-based vocabulary.The input for this mode include a file specify an a regulatory network, a file specify random walk anchor. Loom file contain scRNA-Seq expression matrix(The row are "Gene" and the column are "CellID"). 
3. Gene-only running mode. This mode require only Loom file contain scRNA-Seq expression matrix(The row are "Gene" and the column are "CellID"). 

optional arguments:
  -h, --help            show this help message and exit
  --out OUT             Path for output file
  --n_lda N_LDA         The number of topics LDA model
  --walk_len WALK_LEN   The max length of each random walk
  --n_walks N_WALKS     The max number of walks per anchor
  --random_seed RANDOM_SEED
                        The random seed
  --max_cores MAX_CORES
                        Max core should be ultilized, -1 == use all
  --job_title JOB_TITLE
                        Job name
  --expression_word_count_scale EXPRESSION_WORD_COUNT_SCALE
                        Parameters for gene expression word count
                        scale(default : 100(expression in UMI count))

Command example:
python INSISTC_main.py --network example_network_mini_Human.txt --loom example.loom --anchor example_anchor_Human.txt --n_lda 30 --running_mode 0
python INSISTC_main.py --network example_network_mini_Human.txt --loom example.loom --anchor example_anchor_Human.txt --n_lda 30 --running_mode 1
python INSISTC_main.py --loom example.loom --n_lda 30 --running_mode 2

The example_network_mini_Human.txt and example_network_mini_Mouse should only be used in testing whether the dependencies of INSISTC is correctly installed. To use INSISTC for cell type identification, please use example_network_Human.txt or example_network_mini_Mouse.txt or other Gene regulatory network spefication with same format.