9. Understanding the file formats
The generated networks
The gene correlation networks are created from 2 files. The files are generated at runtime and stored under the "Networks" folder inside the "Resources" folder. These files are created when the user confirms a grouping and presses "generate networks". The user will then have divided some or all cells in a graph into one or more groups. Each group has a unique color.
The NwkCentroids.cnt file contains information about how many networks were created and where in the graph they came from. The file has the following format:
x | y | z | color | graph |
x | y | z | color | graph |
... |
Where x y and z are coordinates for the mean position of the group, color is the color as hexadecimal RGB values, and graph is the name of the graph that it came from. Graphs are always named after the files they originated from.
The Networks.nwk file contains information about which genes are correlated to which. The file format is the following:
pcor | node_1 | node_2 | pval | qval | prob | color | key_1 | key_2 |
pcor | node_1 | node_2 | pval | qval | prob | color | key_1 | key_2 |
... |
Where pcor is the partial correlation. node_1 and node_2 are names of 2 cells that are correlated, color is the color in hexadecimal RGB values of the group this pair of genes belongs to, key_1 and key_2 are simply the two permutations of the two gene names concatenated together. The rest of the columns are not used by CellexalVR. A gene may be featured on more than one row and thus correlated with more than one other gene. Since all groups' networks are in this file, the same pair of genes may featured on more than one row, but in a different group.