Closed StevenCannon-USDA closed 1 month ago
@StevenCannon-USDA not %100 sure but I think your test list may only be enriched for molecular_function (which is how you selected them). I just did a keyword search for "membrane" and the resulting gene list seemed to have enrichment in all three "namespaces". Let us know if you think I'm misunderstanding/misinterpreting
@adf-ncgr Ah - thanks. I'll do some more testing and report.
@adf-ncgr - Thanks for the quick feedback and helpful hint. You are right. My test set just happened to be specific for molecular_function.
Here are three other test sets that work across all three GO aspects. (I am working on this in the process of writing a page to introduce this kind of analysis. It might become a blog post.)
Aspect: Molecular Function
GO:0005215 "transporter activity"
glyma.Wm82.gnm4.ann1.Glyma.01G022700
glyma.Wm82.gnm4.ann1.Glyma.01G035000
glyma.Wm82.gnm4.ann1.Glyma.01G041400
glyma.Wm82.gnm4.ann1.Glyma.01G041450
glyma.Wm82.gnm4.ann1.Glyma.01G042100
glyma.Wm82.gnm4.ann1.Glyma.01G081600
glyma.Wm82.gnm4.ann1.Glyma.01G081700
glyma.Wm82.gnm4.ann1.Glyma.01G105000
glyma.Wm82.gnm4.ann1.Glyma.01G112500
glyma.Wm82.gnm4.ann1.Glyma.01G113400
biological_process: transport [GO:0006810] 3.880204e-12 10
cellular_component: membrane [GO:0016020] 1.894009e-4 10
molecular_function: transporter activity [GO:0005215] 1.600475e-15 10
Aspect: Cellular Component
GO:0005856 "cytoskeleton"
glyma.Wm82.gnm4.ann1.Glyma.01G128700
glyma.Wm82.gnm4.ann1.Glyma.01G155300
glyma.Wm82.gnm4.ann1.Glyma.03G041600
glyma.Wm82.gnm4.ann1.Glyma.03G146200
glyma.Wm82.gnm4.ann1.Glyma.05G158300
glyma.Wm82.gnm4.ann1.Glyma.08G116000
glyma.Wm82.gnm4.ann1.Glyma.09G194900
glyma.Wm82.gnm4.ann1.Glyma.09G279900
glyma.Wm82.gnm4.ann1.Glyma.11G089400
glyma.Wm82.gnm4.ann1.Glyma.13G162500
biological_process: actin cytoskeleton organization [GO:0030036] 4.626113e-10 5
cellular_component: cytoskeleton [GO:0005856] 2.640981e-21 10
molecular_function: microtubule motor activity [GO:0003777] 0.002433 3
microtubule binding [GO:0008017] 0.004066 3
actin binding [GO:0003779] 0.041614 2
Aspect: Biological Process
GO:0007165 "signal transduction"
glyma.Wm82.gnm4.ann1.Glyma.01G032400
glyma.Wm82.gnm4.ann1.Glyma.01G032900
glyma.Wm82.gnm4.ann1.Glyma.01G033200
glyma.Wm82.gnm4.ann1.Glyma.01G033300
glyma.Wm82.gnm4.ann1.Glyma.01G039000
glyma.Wm82.gnm4.ann1.Glyma.01G046900
glyma.Wm82.gnm4.ann1.Glyma.01G060300
glyma.Wm82.gnm4.ann1.Glyma.01G112200
glyma.Wm82.gnm4.ann1.Glyma.01G112300
glyma.Wm82.gnm4.ann1.Glyma.01G125300
biological_process: defense response [GO:0006952] 3.762962e-13 9
signal transduction [GO:0007165] 1.688834e-12 9
cellular_component: No enrichment
molecular_function: ADP binding [GO:0043531] 5.387080e-15 9
protein binding [GO:0005515] 5.267104e-4 9
+1 to writing up some user guidance about the existence of this function- as I recall it had come up as a use case of interest when we had our discussion with Anna Locke.
The Gene Ontology Enrichment widget at https://mines.legumeinfo.org/glycinemine seems to be working for
molecular_function
but notcellular_component
orbiological_process
To test, I enter this list -- which should show enrichment, since all of these genes share the GO terms GO:0005515,GO:0003676 Stabilizer of iron transporter SufD / Polynucleotidyl transferase
Then I save this list of ten genes and check the widget at the bottom of the report page. For
cellular_component
andbiological_process
, the result is "No enrichment found"; whereas for molecular_function, "nucleic acid binding" and "protein binding" are reported as being enriched. I see the same result if I generate a list of all 52K genes in this annotation set and set that as the background population -- so I think the "Default" is probably correct.If I check the report for a single gene ... https://mines.legumeinfo.org/glycinemine/gene:glyma.Wm82.gnm4.ann1.Glyma.02G002500 I see the expected GO terms (GO:0005515,GO:0003676).
Checking one of these GO terms ... https://mines.legumeinfo.org/glycinemine/goterm:GO:0005515 I see that the Namespace is
molecular_function
So, I wonder if it is possible to associate additional namespaces for the GO terms? Or some other method for getting the enrichment widget to know about
cellular_component
andbiological_process
for GO terms?Not urgent, since this is a somewhat specialized function; however, GO enrichment is a useful kind of analysis, and it seems not to be working correctly.