Step0: KNN-graph construction

from sklearn.neighbors import kneighbors_graph
import numpy as np
import pandas as pd
import datetime


# Hyperparameters
InputFolderName = "./MERFISH_Brain_KNNgraph_Input/"
KNN_K = 69

# Import region name list.
Region_filename = InputFolderName + "ImageNameList.txt"
region_name_list = pd.read_csv(
     Region_filename,
     sep="\t",  # tab-separated
     header=None,  # no heading row
     names=["Image"],  # set our own names for the columns
 )

for graph_index in range(0, len(region_name_list)):

 print(graph_index)
 # Import target graph x/y coordinates.
 region_name = region_name_list.Image[graph_index]
 GraphCoord_filename = InputFolderName + region_name + "_Coordinates.txt"
 x_y_coordinates = np.loadtxt(GraphCoord_filename, dtype='float', delimiter="\t")

 K = KNN_K
 print(datetime.datetime.now().strftime('%Y-%m-%d %H:%M:%S'))
 print('Constructing KNN graph...')
 KNNgraph_sparse = kneighbors_graph(x_y_coordinates, K, mode='connectivity', include_self=False, n_jobs=-1)  #should NOT include itself as a nearest neighbor. Checked. "-1" means using all available cores.
 print(datetime.datetime.now().strftime('%Y-%m-%d %H:%M:%S'))
 KNNgraph_AdjMat = KNNgraph_sparse.toarray()
 # Make the graph to undirected.
 print('Making KNN graph bidirectional...')
 KNNgraph_AdjMat_fix = KNNgraph_AdjMat + KNNgraph_AdjMat.T  #2min and cost one hundred memory.
 # Extract and write the edge index of the undirected graph.
 KNNgraph_EdgeIndex = np.argwhere(KNNgraph_AdjMat_fix > 0)  #1min
 filename0 = InputFolderName + region_name + "_EdgeIndex.txt"
 np.savetxt(filename0, KNNgraph_EdgeIndex, delimiter='\t', fmt='%i')  #save as integers. Checked the bidirectional edges.
 print(datetime.datetime.now().strftime('%Y-%m-%d %H:%M:%S'))

Hyperparameters

KNN_K: The K value in the KNN gragh.

Input

The example input data to the unsupervised learning mode of CytoCommunity is a KNN graph based on mouse brain MERFISH data, including cell type labels, cell spatial coordinates, edge index, gragh index and node attributes files and an image name list.

── MERFISH brain KNNgraph input data
   ├─ ImageName_CellTypeLabel.txt
   ├─ ImageName_Coordinates.txt
   ├─ ImageName_NodeAttr.txt
   └─ ImageName_GraghIndex.txt

Output

We use Step0 to construct KNN graghs and prepare data for the subsequent steps, and this step produces “EdgeIndex.txt” file for each input image.

── MERFISH brain KNNgraph input data
   ├─ ImageName_CellTypeLabel.txt
   ├─ ImageName_Coordinates.txt
   ├─ ImageName_NodeAttr.txt
   ├─ ImageName_GraghIndex.txt
   └─ ImageName_EdgeIndex.txt