Input and output files description

Input files

meta data file

A example meta data file stereo_input.csv is provided in Zenodo Dataset repository under preprocessing directory. The meta data file looks like the following:

Cell_ID	Sample	Cell_Type	x	y
E12_E1S3_100034	E12_E1S3	Fibro	15940	18584
E12_E1S3_100035	E12_E1S3	Fibro	15942	18623
…	…	…	…	…
E16_E2S7_326412	E16_E2S7	Fibro	32990.5	14475

• Cell_ID

  The name of each cell. It should be Cell_ID for cell-level data and Spot_ID for low resolution data. Warning: Duplicated Cell_IDs within the same sample are not permitted. In the event of duplicated Cell_IDs across samples, the sample name will be prefixed to Cell_ID.

• Sample

  The name of sample which each cell belongs to.

• Cell_Type

  Cell type for each cell. This column is not required for low resolution data.

• x

  X coordinate for each cell.

• y

  Y coordinate for each cell.

• z (optional)

  Z coordinate for each cell.

Output files

A example output is provided in Zenodo Dataset repository under ONTraC_output directory.

NN-dir

Previouly named as preprocessing-dir.

• meta_data.csv.gz

  Processed meta data file.

• samples.yaml

  File contains the required files information for GNN training.

• {sample name}_CellTypeComposition.csv.gz

  Files contain the cell type composition information for each niche.

• {sample name}_Coordinates.csv

  Files contain the spatial information of anchoring cell for each niche.

• {sample name}_EdgeIndex.csv.gz

  Files contain the niche index of edges among niche graph.

• {sample name}_NeighborIndicesMatrix.csv.gz

  Files contain the neighborhood index of each niche for niche graph.

• {sample name}_NicheWeightMatrix.npz

  Files contain the weights between cells and niches.

• cell_type_code.csv

  File contains the mapping of cell type name to integer.

• spotxcelltype.csv.gz

  Deconvolution methods outputed cell type composition for each spot. This file doesn’t exist when using cell level dataset as input.

GNN-dir

• cell_level_niche_cluster.csv.gz

  File contains the probabilistic assignment of a cell to niche cluster.

• cell_level_max_niche_cluster.csv.gz

  File contains the niche cluster with maximum probability to each cell.

• niche_level_niche_cluster.csv.gz

  File contains the probabilistic assignment of a niche to niche cluster.

• niche_level_max_niche_cluster.csv.gz

  File contains the niche cluster with maximum probability to each niche.

• niche_hidden_features.csv.gz

  File contains the tranined embeddings of each niche.

• {sample name}_out.csv.gz

  Files contain the features for each niche cluster in each sample.

• {sample name}_out_adj.csv.gz

  Files contain the adjancy information between niche clusters in each sample.

• {sample name}_s.csv.gz

  Files contain the projection probabilities from niche to niche clusters in each sample.

• {sample name}_z.csv.gz

  Files contain the embeddings of each niche in each sample.

• consolidate_out.csv.gz

  File contains the features for each niche cluster.

• consolidate_out_adj.csv.gz

  File contains the adjancy information between niche clusters.

• model_state_dict.pt

  File contains the trained parameters for model.

• epoch_0.pt

  File contains the initial parameters for model.

• epoch_X.pt

  File contains the intermediate parameters for model.

NT-dir

Previouly named as NTScore-dir.

• {sample name}_NTScore.csv.gz

  Files contain the niche- and cell-level NT score for each niche/cell.

• NTScore.csv.gz

  File contains niche- and cell-level NT score for all samples.

• niche_cluster_score.csv.gz

  File contains NT score for each niche cluster.