Pytorch Geometric - Build a dataset from multiple graphs with common nodes

I have 2 different graphs, in the form of edge lists, that I would like to integrate into one large graph before training my model on node classification.
These 2 graphs share nodes which correspond to common entities.
How is it possible to integrate the graphs, while preserving information about common nodes?

For instance, let’s consider the following edge lists:

[0 - 1,
5 - 4,
2 -16,
13 - 8]


[5 -12,
15 - 3,
13 - 2]

The first graph edge list has 8 nodes and the second has 6 nodes.
Providing the edge lists as such to the Data object leads to an “out of bound” error during training because the model tries to access nodes according to their index in the edge list. For instance, edge “15-3” makes the model choose the 15th and the 3rd node, however the data object has only 6 nodes, so it returns an “out of bound” error.

Any help or recommendation would be much appreciated! Thank you.