CL pipeline failure - Githubissues

Cellular-Semantics / CellMark

1 stars 1 forks source link

Closed Caroline-99 closed 1 day ago

Caroline-99 commented 2 weeks ago

aleixpuigb commented 4 days ago

We should only add the CL_agreed TRUE, otherwise we might encounter more CL label repetitions.

hkir-dev commented 1 day ago

In the HLCA source file there are two clusters annotated as bronchial goblet cell and CL_agreed TRUE (line 27 and 29), that seems to be the reason.

Is this a copy error made by me, should I keep only one of these?

Also, David mentioned:

If markers are going into CL there needs to be a single cluster per dataset + the CL mapping could differ from CxG.

aleixpuigb commented 1 day ago

One is a more specific term that is not CL agreed.

hkir-dev commented 1 day ago

Sorry, I couldn't understand the solution. Should I make the first one False?

index	cl_class	cl_label	Cell_type	Marker_set	Minimal_markers	Minimal_markers_label	Organ	...	CL_agreed
27	CL:1000312	bronchial goblet cell	Goblet (bronchial)	CLM:1000026	ensembl:ENSG00000197943\|ensembl:ENSG00000174501	PLCG2\|ANKRD36C	Lung	...	True
29	CL:1000312	bronchial goblet cell	Goblet (subsegmental)	CLM:1000028	ensembl:ENSG00000255823\|ensembl:ENSG00000149021	MTRNR2L8\|SCGB1A1	Lung	...	True

aleixpuigb commented 1 day ago

I think this is because the problem with the PR incorporating commits from a different branch. CLM:1000028 should have remained False.

hkir-dev commented 1 day ago