branch lengths missing for standard non-parametric bootstrap contree file

[ti...@posteo.de]

Hi.
Not sure if this is a feature request or a bug.

Running these:
1. iqtree-omp -s example.phy -nt 4 -bb 1000 -pre example.phy.bb1000
2. iqtree-omp -s example.phy -nt 4 -b 100 -pre example.phy.b100

I get these contrees respectively:
1. (LngfishAu:0.1709694637,(LngfishSA:0.1883831199,LngfishAf:0.1648671706)100:0.107347,(Frog:0.2562447908,((((Turtle:0.2215285098,(Crocodile:0.3056817258,Bird:0.2310266099)97:0.065125)71:0.036477,Sphenodon:0.3446959619)49:0.020433,Lizard:0.3860705950)99:0.074003,(((Human:0.1851977105,(Seal:0.0944428932,(Cow:0.0823070449,Whale:0.1013098304)99:0.040444)72:0.025258)90:0.034061,(Mouse:0.0584277962,Rat:0.0905653375)100:0.121824)100:0.060722,(Platypus:0.1920402609,Opossum:0.1510480052)98:0.037331)100:0.149069)100:0.127576)100:0.094066);
2. (Bird,Crocodile,(Turtle,(Sphenodon,(Lizard,((((LngfishAf,LngfishSA)100,LngfishAu)100,Frog)100,((Opossum,Platypus)96,((((Cow,Whale)100,Seal)64,Human)91,(Mouse,Rat)100)100)100)100)38)70)97);

The contree for the standard bootstrap is missing branch lengths, would it be possible to have these?

Thanks.
Tim

[Bui Quang Minh]

Dear Tim,

This is a feature request ;-) I understand that one would like to compute branch lengths of the consensus tree, which IQ-TREE could have done actually. As a workaround for now you can input the contree with -te option via a second run like this:

iqtree-omp -s example.phy -nt 4 -te example.phy.b100.contree

I would also add -m option with the model determined before to save the model selection step.

Cheers, Minh

--
You received this message because you are subscribed to the Google Groups "IQ-TREE" group.
To unsubscribe from this group and stop receiving emails from it, send an email to iqtree+un...@googlegroups.com.
To post to this group, send email to iqt...@googlegroups.com.
Visit this group at https://groups.google.com/group/iqtree.
For more options, visit https://groups.google.com/d/optout.

--

Bui Quang Minh

Center for Integrative Bioinformatics Vienna (CIBIV)

Campus Vienna Biocenter 5, VBC5, Ebene 1

A-1030 Vienna, Austria

Phone: ++43 1 4277 74326

Email: minh.bui (AT) univie.ac.at

[ti...@posteo.de]

perfect, thanks

[Wannes Dermauw]

Hi 

I was wondering whether this is still a feature request (calculation of branch lengths for standard non-parametric bootstrap contree?).

I ran 20 jobs:

"iqtree -s infile.txt -bo 10 -st AA -m LG+G4+I+F -pre bootX -nt AUTO"

then combined 20 "bootX.boottrees" to a file named "alltrees" (using cat in linux)

and then used the following command to create contree:

"iqtree -con -t alltrees"

but the contree is without branch lengths ...

what would be the best way to include branch lengths?

regards

Wannes

Op donderdag 28 september 2017 14:28:55 UTC+2 schreef ti...@posteo.de:

[Minh Bui]

Hi Wannes,

You can input the consensus tree using -te option with the original alignment, e.g:

iqtree -s ALN_FILE -te CONTREE_FILE 

to obtain its branch lengths estimated using the original alignment.

I have added this into the FAQ.

Thanks

Minh

To view this discussion on the web visit https://groups.google.com/d/msgid/iqtree/f68a5232-4d84-47ec-9022-e315afbd5cad%40googlegroups.com.

[Minh Bui]

Hi Wannes,

The key question here is: what is your defition of “well-supported" branches? 

The two methods, UFBoot and SBS, have different meanings for support values, see: 

http://www.iqtree.org/doc/Frequently-Asked-Questions#how-do-i-interpret-ultrafast-bootstrap-ufboot-support-values

I’d only start to “believe” in a clade with UFBoot >= 95%, whereas this threshold for SBS is lower, say 70 or 80%.

However, note that for phylogenomic (multi-gene) data, both values tend to be 100%, which is not always trustworthy. In that case, you can perform concordance factor analysis:

http://www.iqtree.org/doc/Concordance-Factor

that complements the bootstrap approach.

Cheers

Minh

PS: I replied to the IQ-TREE group to keep track of the answers, and this also might be of general interests.

On 31 May 2020, at 7:27 pm, Wannes Dermauw <dermau...@gmail.com> wrote:

Dear Minh

Thank you for the quick reply suggested solution.

I do have a more general question regarding consensus trees obtained through ‘classical” bootstrapping versus consensus tree obtained through “UFboot” bootstrapping.

For the same input alignment, I created a consensus tree

-using either 3000 UFboot bootstraps (ran on CIPRES server): UFBOOTtree

or

-200 “classical” (SBS) bootstraps (split into 20 jobs of 10 bootstraps, could not do more jobs because of limited use of CPU/user). SBSboottree

However, somel well-supported branches in the UFBOOTtree are not conserved in the SBSboottree. Hence, my question is, which consensus tree should we “believe” in?

The one based on a high number of rapid bootstrapping (UFboot) or the one based on a low number of classical bootstrapping (SBS)?

Kind regards

Wannes