Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
  • May 25 08:16
    anatshafir1 commented #160
  • May 24 15:19
    bredelings commented #160
  • May 24 13:07
    bredelings commented #160
  • May 24 13:06
    bredelings commented #160
  • May 24 12:56
    anatshafir1 commented #160
  • May 24 10:54
    bredelings commented #160
  • May 24 09:22
    anatshafir1 opened #160
  • May 18 21:16

    bredelings on simpler-rollback2

    Add SolutionRollbackInfo parame… Add a const Record all the rollback info in… (compare)

  • May 18 20:40

    bredelings on simpler-rollback2

    Factor out Merge( ) into its ow… Factor out Assign( ). Factor failure code out into Ma… and 4 more (compare)

  • May 18 19:31

    bredelings on simpler-rollback

    Make the SolutionRollbackInfo m… Record how many times each solu… Factor out RemoveImpliedSplits(… and 7 more (compare)

  • May 18 16:22

    bredelings on python3

    (compare)

  • May 18 15:51

    bredelings on development

    Forward leaf ottids -- maybe no… (compare)

  • May 12 23:50
  • May 01 08:37
  • Apr 25 17:52

    jimallman on pyramid-migration

    Refactor "all-types" push and m… (compare)

  • Apr 25 17:50

    jimallman on pyramid-migration

    Allow test access directly to :… (compare)

  • Apr 25 17:49

    jimallman on pyramid-migration

    Restore restrict access (localh… (compare)

  • Apr 23 05:48

    jimallman on pyramid-migration

    Allow web access through any IP… (compare)

  • Apr 23 05:25

    jimallman on pyramid-migration

    Use standard web port for Ansib… (compare)

  • Apr 23 04:37

    jimallman on pyramid-migration

    Use 'git+https' vs 'git+ssh' fo… (compare)

nclaidiere
@nclaidiere

Hi, I posted this on github but this is probably a better place for it. I am new to rotl and just trying to run the getting started tutorial...

I installed rotl package on Rstudio and just did what was in the tuto:
library(rotl)
apes <- c("Pongo", "Pan", "Gorilla", "Hoolock", "Homo")
(resolved_names <- tnrs_match_names(apes))

This returned the following error:
Error: HTTP failure:500
Error: could not connect to http://localhost:1984/v3/tnrs/contexts

I tried shutting down the firewall just in case but to no avail. I am on windows, running the latest R and Rstudio versions. Nothing on the forums or here as far as I can tell.

Any ideas?

Benjamin Redelings
@bredelings
A server process appears to have crashed -- I restarted it and it works fine now.
Benjamin Redelings
@bredelings
^ @nclaidiere
I think I see the bug -- should be able to fix this so it doesn't happen again.
Benjamin Redelings
@bredelings
OK, fixed.
nclaidiere
@nclaidiere
Yes! great! it works just fine. Thanks.
Jim Allman
@jimallman
hi @all! Does anyone know why a login is now required for all TreeBASE URLs? This is blocking our ability to import studies from TreeBASE, at least in my testing:
https://treebase.org/
https://treebase.org/treebase-web/
https://treebase.org/treebase-web/phylows/study/TB2:S12586?format=nexml
Jim Allman
@jimallman
(confirmed, create-study-from-treebase-id seems to be broken on production as well)
Emily Jane McTavish
@snacktavish
eep, is it because Luna and I hammered them too hard downloading data for physcraper?
I will send some emails and investigate
Jim Allman
@jimallman
thanks! but i don't think it's personal... i'm sending queries direct from my desktop (and trying URLs in a web browser) and all are blocked by a login prompt
Emily Jane McTavish
@snacktavish
Oh ya, I was thinking we could have inspired a more general lock down! I tweeted, but not responses
terry06890
@terry06890

Hi, I'm not sure if I'm posting in the right place. Apologies if not.

I'm looking for clarification about the online tree browser. I understand that paths can be solid or dotted, representing whether they are "supported by phylogeny", or "supported only by taxonomy". But I'm unclear on how this is determined from the data in the synthesis release.

I know annotations.json contains "supported_by" and "conflicts_with" information on each node, and expected "supported by phylogeny" perhaps meant there was at least one supporting tree, or more supporting trees than conflicting trees, but neither seems true. For instance, with mrcaott100410ott205763, it's displayed as supported by phylogeny, but it has 3 conflicting trees, with only 1 supporting tree. And, with mrcaott100399ott100410, it's displayed as "supported only by taxonomy", but it has 5 supporting trees and no conflicting trees.

Is there something obvious that I'm missing?

Benjamin Redelings
@bredelings
For mrcaott100399ott100410, I'm seeing it with solid lines and 3 supporting trees + 1 conflicting tree, which seems to be different from what you are seeing...: https://tree.opentreeoflife.org/opentree/argus/opentree13.4@mrcaott100399ott100410
For mrcaott100410ott205763, I see solid lines with 1 supporting tree + 3 conflicting trees, as you say: https://tree.opentreeoflife.org/opentree/argus/opentree13.4@mrcaott100410ott205763
I think maybe you have two questions:
  • Q1: what do the solid/dashed lines mean?
  • Q2: when there is conflict between the input phylogenies about which branches to include, how do we decide which branches make it into the synthesis tree?
It might be worth taking a look at the paper for the supertree approach: https://peerj.com/articles/3058/
Benjamin Redelings
@bredelings
So, for Q1: the solid lines mean that there is at least one input phylogeny that supports the branch. It isn't a measure of the number of phylogenies for-or-against the branch. Its about whether the branch is taxonomy-only.
For Q2: the input phylogenies are ranked. We use the rank order rather than the number of trees for and against to decide which branches make it into the synthesis supertree.
So, I suspect that the single tree supporting mrcaott100410ott205763 is ranked higher than the tree trees that conflict with it.
See https://tree.opentreeoflife.org/curator/collection/ which shows the "collections". A collection is just an ordered list of trees.
Benjamin Redelings
@bredelings
Not all collections are used to make the synthesis tree... I'd have to check to see which collection contains the supporting tree ot_951 and the conflicting trees ot_1091 etc.
Hmm... this almost seems like it should be written down as an FAQ somewhere...
terry06890
@terry06890

Thanks Ben, that clarifies a lot. Regarding mrcaott100399ott100410, I checked again, and am seeing what you saw. I probably mixed it up with another node.

Out of curiosity, I had a closer look at mrcaott100410ott205763 (solid line, 1 supporting, 3 conflicting):

  • ot_951 is supportive, and is in collection 'Reef Fishes' with rank 4
  • ot_1091 is conflicting, and is in collection 'Actinopterygii' with rank 5, and in 'Reef Fishes' with rank 14
  • ot_1592 is conflicting, and is in collection 'Inputs to synthesis (default)' with rank 370
  • pg_2651 is conflicting, and is in collection 'Inputs to synthesis (default)' with rank 71, and in 'Actinopterygii' with rank 1, and in 'Reef Fishes' with rank 12

The supportive tree ot_951 outranks ot_1091 and pg_2651, via the 'Reef Fishes' collection, but doesn't share a collection with ot_1592. Perhaps the ranks are compared across collections, so the rank 4 takes priority over the 370. pg_2651 does have a rank 1 though, and it's unclear how this affects things.

I think I'll let it go here. I had a look at the paper you mentioned, and was fairly overwhelmed. I wasn't intending to go that deep. Your response was helpful enough for what I'm trying to do. Thank you.

Benjamin Redelings
@bredelings
The collections go into synthesis in a certain order too, so a tree in a higher-ranked collection will be higher-ranked than a tree in a lower-ranked collection. I haven't looked through all the collection ranks to figure out if including ot_951 in the synth tree makes sense though.
Yeah, the paper is complex. With papers like that, you can usually get the big idea by skipping the mathy parts and reading the section headings + some of the non-mathy parts. I wouldn't recommend reading it linearly from start-to-finish, unless your goal is to implement a similar method (which it is not).
In this case, I think the 4 goals + 1 non-goal could give helpful background.
terry06890
@terry06890
I found something else rather strange. It seems there are two nodes named "Eutheria", but one seems to be erroneous.
Benjamin Redelings
@bredelings
This is going to be a taxonomy issue. Our taxonomy (OTT = OpenTree Taxonomy) is created by combining information from several other taxonomies. Sometimes the process goes wrong, but in this case it looks OK.
Sometimes taxonomic names occur at multiple positions in the tree, so having Eutheria occur multiple times is not necessarily wrong.
If you click on the GBIF and IRMNG links under "Source taxonomy", you'll see that other taxonomies also say that there is a Eutheria that is in Coleoptera, and that it goes back to 1909.
So, in this case, it is probably not a mistake that was made which merging the taxonomies. There definitely are mistakes though.
terry06890
@terry06890
Ah okay, that makes sense. Thanks for clearing that up.
TonyRees
@TonyRees
Thanks for this alert... what we have is an example of a cross rank homonym, not very well supported in IRMNG at this time except for names at major ranks i.e. kingdom/phylum/class/order/family/genus, although I have just added the mammal instance of "Eutheria" (Eutheria Huxley, 1880, apparently a different concept from Eutheria Gill, 1872, which included both placental mammals and marsupials) at the level of infraclass, per Ruggiero et al.'s 2015 treatment (new record at https://www.irmng.org/aphia.php?p=taxdetails&id=11922425). Accordingly I am also adding Metatheria to IRMNG as the other infraclass of mammals, as well as Theria as a subclass... previously I just jumped straight from class to order, there are always intervening things that can be added :) Also just noted in passing that an order of mammals, Cingulata Illiger, 1811, has several cross rank homonyms at genus level (Cingulata Méhelÿ, 1935 in Hymenoptera, Cingulata Laseron, 1959 in Mollusca) so there will plenty of other cross rank homonym examples lurking, to be sure.
TonyRees
@TonyRees
The above of course points out the need to disambiguate homonyms, which can be assisted by the cited authority (not quite a unique identifier in conjunction with the name, but nearly always so), the type of organism concerned (again some wrinkles here, as some things get moved around or are of uncertain placement), and the rank of the name (especially if quoted ranks are different)...
Emily Jane McTavish
@snacktavish
We disambiguate by type of organism, in a big picture taxonomy sense - https://tree.opentreeoflife.org/taxonomy/browse?id=683263 is labelled "Eutheria (in Deuterostomia)" and https://tree.opentreeoflife.org/taxonomy/browse?id=3373561 is "Eutheria (genus in Protostomia)". Keeps me on my game with regard to deep taxonomy!
The parens in our taxon names can generate problems elsewhere though, unfortunately.
terry06890
@terry06890
Thanks for the additional clarification. I appreciate it :)
TonyRees
@TonyRees
@snacktavish wrote: /We disambiguate by type of organism/ OK, will work most of the time, except when cross rank homonyms are present in the same "organism type" - may be a rare event though :) (if you are going to ask me for examples, I cannot think of any off hand, but do vaguely recall encountering the odd one - will advise if I find any!!)
Emily Jane McTavish
@snacktavish
@TonyRees We finesse 'organism type' to the degree required for disambiguation! So as long as they aren't sister taxa (and that wouldn't happen) we can append whatever taxonomic info is needed.
TonyRees
@TonyRees
OK, sounds good.
You might run into some problems regarding taxon concepts though - for example same name, same author, but sens. lat in some treatments, sens. strict. in others (certain children excluded)... not an area I have worried about too much (I just pick one or the other at a particular time, and say whom I am following) but for taxomomy merging actions could be a bit tricky. But let me know if I am straying too far off topic...
Jim Allman
@jimallman
FYI to @all, i'm testing the new phylesystem-api server (ot62) from devtree. Currently sorting out some filesystem permission problems that are breaking the curation app. Apologies for any inconvenience, and let me know if I should restore normal operations.
jimallman @jimallman is still tracking down needed Ansible variables (some missing, some located in other YAML files)
Yan Wong
@hyanwong
Why is it that most papers will happily publish their sequence data (e.g. in Genbank), and make their phylogenies available as bitmaps or vector PDFs, but not bother making the newick or nexus format for their trees available? I've never understood this. It has made it basically impossible to store many trees from studies in a sensible form.
Emily Jane McTavish
@snacktavish
@hyanwong I ask myself that question nearly every day!! :sob:
Yan Wong
@hyanwong
Still, AI is getting to the point where it should be possible to extract the data from the figures, right? right???