Potential bug in panaroo

[Hocnonsense]

I'm reading the code of panaroo, and I've found something difficult to understand

One is in cdhit.py, is it a typo for "as=AS"? it may be

            run_cdhit(
                ...
                aL=aL,
                AL=AL,
                aS=aS,
                AS=AS,
                ...
            )

in

panaroo/panaroo/cdhit.py

Lines 186 to 210 in e928a7a

	if dna:
	run_cdhit_est(input_file=temp_input_file.name,
	output_file=temp_output_file.name,
	id=id,
	s=s,
	aL=aL,
	AL=AL,
	aS=AS,
	accurate=accurate,
	use_local=use_local,
	strand=strand,
	quiet=quiet,
	n_cpu=n_cpu)
	else:
	run_cdhit(input_file=temp_input_file.name,
	output_file=temp_output_file.name,
	id=id,
	s=s,
	aL=aL,
	AL=AL,
	aS=AS,
	accurate=accurate,
	use_local=use_local,
	quiet=quiet,
	n_cpu=n_cpu)

Another is in generate_network.py, I think genes at the end of contigs should be marked as node['hasEnd']=True

panaroo/panaroo/generate_network.py

Lines 69 to 78 in e928a7a

	for i, id in enumerate(seq_ids):
	current_cluster = seq_to_cluster[id]
	seqid_to_centroid[id] = cluster_centroids[current_cluster]
	loc = id.split("_")
	genome_id = int(loc[0])
	if loc[-1] == "0":
	# we're at the start of a contig
	if prev is not None: G.nodes[prev]['hasEnd'] = True
	prev = current_cluster
	if G.has_node(prev) and (prev not in paralogs):

However, at the end of cycle, the last gene was ignored.

panaroo/panaroo/generate_network.py

Lines 226 to 228 in e928a7a

	prev = current_cluster
	return G, centroid_context, seqid_to_centroid

Is it necessary to add another check for it such as:

    if prev is not None:  # end of cycle
        G.nodes[prev]["hasEnd"] = True

    return G, centroid_context, seqid_to_centroid

Thanks if you could reply!

[gtonkinhill]

Hi,

Sorry for the delayed reply, and thank you very much for flagging these issues.

The first is indeed a typo, which I will fix in the next release. However, it should not affect the output, as Panaroo only uses global alignment (use_local=False), which overrides this parameter.

The second issue appears to be a potential bug, although it would likely have a very small impact on the output. It could result in at most one miscalled gene if an annotation error occurred in the last gene processed by Panaroo. I will investigate this further and fix it if necessary in the next release.