lbcb-sci/gnnome

A framework for training graph neural networks to untangle assembly graphs obtained from OLC-based de novo genome assemblers.

gnnome's Issues

About the low depth sequence assmble

Dear developer:
I am trying to assemble the genome using about 15-20X nanopore data and about 15X illumina genome. The genome we are going to assemble , I would like to ask if this is feasible. I don't know if the software can meet this need?
Best day!

Erro in polish step IndexError: list index out of range

Dear developer:
we used the gnnome well in example data, but what we met the erro is we used the nanopore assemble from raven “raven --threads 100 sample.fastq.gz > sample.raven.contigs.fasta” and we got the fasta file ,then we used the fa2gfa translate the fa to gfa format “fa2gfa -i sample.raven.contigs.fasta -o sample.raven.contig.gfa”.

When we used the step "python /home/test/Software/GNNome-0.4.0/create_inference_graphs.py --reads sample.fastq.gz --gfa sample.raven.contig.gfa --asm raven --out sample.raven" we met the erro：

Starting to parse assembler output
Starting to loop over GFA
Elapsed time: 58s
Elapsed time: 58s
Calculating similarities...
0it [00:00, ?it/s]
Done!
Elapsed time: 58s
Traceback (most recent call last):
File "/home/test/Software/GNNome-0.4.0/create_inference_graphs.py", line 50, in
create_inference_graph(gfa, reads, out, asm)
File "/home/test/Software/GNNome-0.4.0/create_inference_graphs.py", line 13, in create_inference_graph
graph, pred, succ, reads, edges, read_to_node, _ = graph_parser.only_from_gfa(gfa_path, training=False, reads_path=reads_path, get_similarities=True)
File "/home/test/Software/GNNome-0.4.0/graph_parser.py", line 345, in only_from_gfa
graph_dgl = dgl.from_networkx(graph_nx, node_attrs=node_attrs, edge_attrs=edge_attrs)
File "/home/test/anaconda3/envs/gnnome/lib/python3.8/site-packages/dgl/convert.py", line 1345, in from_networkx
g.edata[attr] = F.copy_to(_batcher(attr_dict[attr]), g.device)
File "/home/test/anaconda3/envs/gnnome/lib/python3.8/site-packages/dgl/convert.py", line 1307, in _batcher
if F.is_tensor(lst[0]):
IndexError: list index out of range

Can you help us and tell us how to solve the problem?