search

ekg / seqwish

alignment to variation graph inducer
MIT License
143 stars 18 forks source link

Dead end nodes of 0 length #116

Open casper-schutte opened 11 months ago

casper-schutte commented 11 months ago

I ran seqwish while constructing a graph for validation of the FantasticLamp pipeline and I noticed that around the homology arms I inserted for the edits, there are these weird dead end nodes that have 0 length. In the figure below I am using Bandage to view the graph. There are 3 overlapping edits in the picture, the dead-end nodes are all 0bp in length. All other aspects of the graph are created correctly. I will add the files and output of what I ran in a .zip file. I am using version v0.7.9-0-gd9e7ab5

The seqwish commands are in the create_edit_graph.sh bash script.

![Screenshot from 2023-09-27 15-55-34](https://github.com/ekg/seqwish/assets/4

input_files.zip 7626336/0a0423f3-e0a1-4a2f-96d6-1ae9a055751c)

casper-schutte commented 11 months ago

I managed to find the cause of the problem. In the CSV file where I was extracting the homology arms, the last character on each line was a "Return" character, which showed up as a node with 0 length. I fixed this by changing the way the data is extracted.