Problem with LCA assignment in rma file

User Questions
1

I have converted minimap alignments (alignment to nt database) to rma files using the the sam2rma tool. When I view the rma files in the Megan GUI, I am expecting the LCA (default sam2rma settings) to be based on a top percent value of 10 and min score of 50. I am encountering taxonomic classifications that appear to be inconsistent with the LCA settings. You can see an example of this below.

Screen Shot 2023-08-03 at 11.02.18 AM
Screen Shot 2023-08-03 at 11.02.18 AM1125×630 101 KB

I would expect the classification to have been the LCA of the top taxa (score 1552) and anything within 10% score (> 1397). Instead the read has been classified as a taxon that does not include the taxon of the top hit. I can’t seem to work out the basis on which the LCA seemingly ignored to top hit. The top hit read appears to pass all of the sam2rma settings. I’ve so far only encountered this issue with the top hit taxon (TaxonID: 11086)
Can you please help with this issue?

2

You are right, this is not what the LCA should do (unless you are using the weighted LCA?) If you can give me access to a small file that exhibits this problem, then I will debug this.

3

Great, thanks.
The LCA is long read, % to cover 100.
The example illustrated above is from the following file:

4

Thank you for this interesting example. I agree that this is problematic.

The issue might be that the Negishi virus alignment just sneaks in with a score of 1398. It is a child of Tick-borne encephalitis virus:

image
image2574×1390 434 KB

However, the best assignment in this case would be to Flavivirus.

I will look into this.

5

Thank you for looking into this. I did notice changing percent cover to 51% gives more sensible classifications.

6

Hi Daniel
I am getting further examples of this misclassification in other microbial taxa. I can provide additional examples of these if you are interested. Again, setting percent cover to 51% seems to correct this, but I’m not sure why.