Phylogenetic dating - calibration and confidence intervals

[Tricia Goulding]

Hello,

I am trying the phylogenetic dating feature of IQtree2 (with Windows). I have run a few analyses on a DNA alignment with a starting tree from RAxML, and setting an ancestral date to calibrate the tree. To begin, I tried to calibrate one node with a date file. I made a date file listing the names of 9 taxa and an age for the common ancestor [taxa1,taxa2,taxa3,taxa4,taxa5,taxa6,taxa7,taxa8,taxa9 -20]. 

Command line: iqtree2 -s alignment.phy --date datefile3.txt --date-tip 0 -te RAxML_bestTree_rooted.newick

However, in the output .nex tree file, a different node has this age; the node is 2 nodes basal to the intended node. It seems to me that the wrong node has been calibrated. I have checked the datefile, but it includes only the 9 taxa that descended from the node which I wanted to set as the calibration point. I've re-run the analysis, without a starting tree, but had the same problem with the calibration. In another analysis, I used the same alignment, the same tree, and the same datefile, but added a partition file to the command line. This resulted in a node from another clade having the age from the datefile. This remained the same if I added a second calibration point to the datefile.

iqtree2 -s alignment.phy -p partition.txt --date datefile3.txt --date-tip 0 -te RAxML_bestTree_rooted.newick

Is there a specific format for the date file that could be causing this problem? I have the names separated with commas and a tab separating the taxa from the age of the common ancestor.

The second question I have is about the confidence intervals. In the analyses I've run, the confidence intervals are all one-sided, where I'm used to seeing 2-sided confidence intervals. I used "--date-ci 100" in the command line to get the confidence intervals. For most analyses I tried, the confidence intervals are all older (more negative) than the estimates (tMRCA -113.123 [-155.976; -113.123]) but for a few analyses I had confidence intervals that were younger than the estimates (tMRCA -73.0248 [-73.0248; -39.635]). Are these confidence intervals meant to be centered on the nodes? Or should an interval of equal size be applied to both sides of tMRCA? 

Thanks,

Tricia

[Minh Bui]

Hi Tricia,

Sorry I overlooked your email. I forward to Hien, the author of LSD2, who may have an answer… 

Please update this version https://github.com/Cibiv/IQ-TREE/releases/tag/v2.0.6. Hien has fixed several issues in LSD2. Your date file seems OK. If the problem persists pls let us know.

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 view this discussion on the web visit https://groups.google.com/d/msgid/iqtree/01efc207-bc17-40cf-b3a4-4bf429ef2d9bo%40googlegroups.com.

[Tricia]

Hi Minh and Hien,

Thank you for your responses to my issue. I have downloaded v.2.0.6 and this fixed the issue with the calibration of the node.

I had another issue with not being able to visualize the confidence intervals in FigTree from the analysis I made in IQTree v.2.0.6 (Windows). I opened the nexus file in a text editor and removed all the quotation marks, and then the confidence intervals could be viewed in FigTree. 

Thank you both for your help.

Best regards,

Tricia

Minh

To unsubscribe from this group and stop receiving emails from it, send an email to iqt...@googlegroups.com.