ArnovanHilten
commented
1 year ago Hi @lesyngenta,
It is quite hard to make plots that work/look good for all the networks. If you post your connection_weights.csv I can provide you some code to plot the results. Did you check the other plotting options? You can use sunburst or layer_weight with the corresponding layer that you want to plot. These should also provide you some insight.
About the NPZ:
The GenNet.py topology create_gene_network should provide also the equivalent .npz file for the topology.csv
The npz file is just a handy alternative to the topology.csv
In short:
Dense layer with custom connections. The custom connections are defined by the mask input, a sparse (COO) connectivity matrix.
# The matrix has the shape of (N_nodes_layer_1, N_nodes_layer_2).
# It is a sparse matrix with zeros for no connections and ones if there is a connections. For example.
# output
# 1 2 3 4 5
# input 1 | 1 0 0 0 0 |
# input 2 | 1 1 0 0 0 |
# input 3 | 0 1 0 0 0 |
# input 4 | 0 1 0 0 0 |
# input 5 | 0 0 1 0 0 |
# input 6 | 0 0 0 1 0 |
# input 7 | 0 0 0 1 0 |
# This connects the first two inputs (1,2) to the first neuron in the second layer.
# Connects input 2,3 and 4 to output neuron 2.
# Connects input 5 to output neuron 3
# Connects input 6 and 7 o the 4th neuron in the subsequent layer
# Connects nothing to the 5th neuronIf you have such a matrix in a numpy array you can simply use:
import numpy as np
a=np.eye(3). # your matrix
import scipy.sparse
coo_matrix = scipy.sparse.coo_matrix(a)
scipy.sparse.save_npz(savepath +'/coo_matrix', coo_matrix)
lesyngenta
Hi Arno,
After I got the results 'connection_weights.csv', I don't know how to interpret that. Shall I look at the 'raw importance' column for the genes most correlated with a phenotype? I like the example figure "example_manhattan.png", but the GenNet.py plot only plots manhattan for SNPs instead of genes. One more question is about how to use ".npz" file in the training. Is there any example for the format of npz file?
Thanks, Le