balabanmetin / uDance

BSD 3-Clause "New" or "Revised" License
35 stars 4 forks source link

ASTRAL issue #4

Open jiyideanjiao opened 10 months ago

jiyideanjiao commented 10 months ago

I used the "tree" option with backbone.nwk and the alignments of 12 genes including 2538 species. I always got error message "In cluster species/output/udance/3, 1 gene tree(s) with lower than 0.80 median lpp are discarded. ASTRAL job on partition species/output/udance/3 has failed. Check the log file species/output/udance/3/astral.incremental.log for further information."

When I check the astral.incremental.log file, it shows as below:

11 trees read from species/output/udance/3/astral_input.trees All output trees will be arbitrarily rooted at species1232

Shutting down threading Exception in thread "main" java.lang.RuntimeException: Error: Leptothorax_acervorum was not seen in main input trees. at phylonet.coalescent.TaxonIdentifier.taxonId(TaxonIdentifier.java:38) at phylonet.coalescent.CommandLine.readInputTrees(CommandLine.java:833) at phylonet.coalescent.CommandLine.runInference(CommandLine.java:639) at phylonet.coalescent.CommandLine.process(CommandLine.java:549) at phylonet.coalescent.CommandLineMP.process(CommandLineMP.java:194) at phylonet.coalescent.CommandLineMP.main(CommandLineMP.java:229)"

Do you have any idea to figure out this issue? Thanks

balabanmetin commented 10 months ago

It must be related to the fact that one of twelve gene trees is discarded because it is low quality ("lower than 0.80 median lpp"). The missing species (Leptothorax_acervorum) is probably present in the backbone tree but none of the remaining 11 gene trees. Please confirm by searching this species in the astral_input.trees. If this is the case, you can resolve this by increasing gene filtering threshold to 1.0, which disables gene filtering (all 12 will be used by ASTRAL):

  # 1D k-means-based (k=2) outlier gene detection parameter. Centroid difference must be 
  # larger than this value to designate the first cluster as outliers.
  outlier_difference: 0.1

I am currently on vacation. I am planning on resolving this 7 days later if this does not help.

balabanmetin commented 10 months ago

A bad gene tree can occur due to failed or bad MSA or lack of phylogenetic signal in the gene.

jiyideanjiao commented 10 months ago

Get it. thanks very much!

balabanmetin commented 10 months ago

Is the issue resolved?

jiyideanjiao commented 10 months ago

still got error message at ASTRAL step