Incongruent results from BEAST compared to trees built by other softwares

[Felipesnailbro Lin]

Hello, I'm using BEAST2 to estimate the divergence time of five species, and I tried both node-dating (Yule Model) and tip dating (FBD) methods. Despite the results of time estimation looked convincing for me, the topologies generated by BEAST2 had slightly different for the taxa in one species. More specifically, one of the lineages become basal group for the species in MrBayes, RAxML, and IQtree, while it became the ingroup from the result of BEAST2. 

I tried to set the topology being fixed through the analysis by editing the 'operator', and I also tried to make the groups as monophyletic while excluding the basal group just like the results from  MrBayes, RAxML, and IQtree. However, none of these approaches worked really well since the divergence time jump to 5 to 10 times larger than the original results (also the divergence time for this species from previous study). I wonder if I should keep trying to let the topology fixed like the results from MrBayes, RAxML, and IQtree, or maybe I should find proper explanation about this incongruence?

I appreciate any comments on this issue. Thanks!

[xj zuo]

Hello,I have encountered the same issue as you while performing evolutionary analysis using BEAST 1. Despite achieving an Effective Sample Size (ESS) greater than 200 for my samples, the resulting tree topology are inconsistent with those obtained using other tools such as RAxML, IQ-TREE, and MStree.

To address this discrepancy, I've recently attempted to use the FixedTreeAnalysis package in BEAST 2 to constrain the tree topology.  However, I am unable to input time, which consequently prevents me from calculating Bayesian Factor (BF) values, and I also cannot create maps that depict transmission relationships.

I'd like to inquire whether you have resolved this issue or have insights into the reasons behind this problem. Your expertise and guidance would be greatly appreciated.

I eagerly await your response .Thanks!

[Felipesnailbro Lin]

Hello xj zuo

I eventually regarded the occurrence of the issue due to the different algorithms from each method I applied. Specifically, the incongruence only exists in closely-related (intraspecific) lineages of my focal group, which were reported with taxonomic issues in previous study. If you really intend to use certain topology, here is something I will recommend:

1. provide valid evidence to explain why you use a certain topology

2. use monophyletic setting in the BEAST prior instead throwing in a fixed tree

I think it is ambiguous to give you the opinions since I'm not sure how severe of incongruence in your trees. 

Hope you find this helpful.

Cheers

xj zuo 在 2023年10月20日 星期五凌晨12:34:29 [UTC+8] 的信中寫道：

[xj zuo]

Thanks for answering my question.

I have already used the monophyletic setting in the BEAST prior, but the results remain the same. Upon a thorough examination of the results, I've noticed some structural differences, specifically, a small cluster near the outgroup in the MLtree is positioned behind other branches in the phylogenetic tree.

From my observations, it appears that the MCC tree's structure is sorted by time, whereas both the MStree and MLtree are sorted based on SNP distances. Below is a comparison of the two trees, with darker colors indicating closer proximity to the present time.The circle represents the cluster with a different position, and the dashed circle indicates where this cluster should be according to the MLtree in the BEAST MCCtree.

Thank you very much for your response.

[Farman, Mark L.]

At a first glance, it appears that the numbers of taxa in the two trees are different. In this case, it is totally expected that you will get different topologies when there is gene flow involving one or more omitted taxa. 

In general, one should not use branch positions, orders or lengths to try and infer anything about the evolutionary history of one’s organism of choice unless one already knows its evolutionary history. This discrepancy is a perfect illustration of this principle. Which one is correct? - probably neither. 

It would help to know the organism…

Mark L. Farman 
Professor, Department of Plant Pathology

225 Plant Science Building
1405 Veteran's Dr.
University of Kentucky
Lexington, KY 40546 USA 
tel:  (859) 218-0728 
fax:  (859) 323-1961

Sent from my iPhone

On Nov 29, 2023, at 2:29 AM, xj zuo <xjzu...@gmail.com> wrote:



You don't often get email from xjzu...@gmail.com. Learn why this is important

CAUTION: External Sender

Thanks for answering my question.

I have already used the monophyletic setting in the BEAST prior, but the results remain the same. Upon a thorough examination of the results, I've noticed some structural differences, specifically, a small cluster near the outgroup in the MLtree is positioned behind other branches in the phylogenetic tree.

From my observations, it appears that the MCC tree's structure is sorted by time, whereas both the MStree and MLtree are sorted based on SNP distances. Below is a comparison of the two trees, with darker colors indicating closer proximity to the present time.The circle represents the cluster with a different position, and the dashed circle indicates where this cluster should be according to the MLtree in the BEAST MCCtree.

Thank you very much for your response.

<不正确.jpg>

在2023年11月24日星期五 UTC+8 14:38:12<Felipesnailbro Lin> 写道：

Hello xj zuo

I eventually regarded the occurrence of the issue due to the different algorithms from each method I applied. Specifically, the incongruence only exists in closely-related (intraspecific) lineages of my focal group, which were reported with taxonomic issues in previous study. If you really intend to use certain topology, here is something I will recommend:

1. provide valid evidence to explain why you use a certain topology

2. use monophyletic setting in the BEAST prior instead throwing in a fixed tree

I think it is ambiguous to give you the opinions since I'm not sure how severe of incongruence in your trees. 

Hope you find this helpful.

Cheers

xj zuo 在 2023年10月20日 星期五凌晨12:34:29 [UTC+8] 的信中寫道：

Hello,I have encountered the same issue as you while performing evolutionary analysis using BEAST 1. Despite achieving an Effective Sample Size (ESS) greater than 200 for my samples, the resulting tree topology are inconsistent with those obtained using other tools such as RAxML, IQ-TREE, and MStree.

To address this discrepancy, I've recently attempted to use the FixedTreeAnalysis package in BEAST 2 to constrain the tree topology.  However, I am unable to input time, which consequently prevents me from calculating Bayesian Factor (BF) values, and I also cannot create maps that depict transmission relationships.

I'd like to inquire whether you have resolved this issue or have insights into the reasons behind this problem. Your expertise and guidance would be greatly appreciated.

I eagerly await your response .Thanks!

在2023年6月2日星期五 UTC+8 02:04:59<Felipesnailbro Lin> 写道：

Hello, I'm using BEAST2 to estimate the divergence time of five species, and I tried both node-dating (Yule Model) and tip dating (FBD) methods. Despite the results of time estimation looked convincing for me, the topologies generated by BEAST2 had slightly different for the taxa in one species. More specifically, one of the lineages become basal group for the species in MrBayes, RAxML, and IQtree, while it became the ingroup from the result of BEAST2. 

I tried to set the topology being fixed through the analysis by editing the 'operator', and I also tried to make the groups as monophyletic while excluding the basal group just like the results from  MrBayes, RAxML, and IQtree. However, none of these approaches worked really well since the divergence time jump to 5 to 10 times larger than the original results (also the divergence time for this species from previous study). I wonder if I should keep trying to let the topology fixed like the results from MrBayes, RAxML, and IQtree, or maybe I should find proper explanation about this incongruence?

I appreciate any comments on this issue. Thanks!

--
You received this message because you are subscribed to the Google Groups "beast-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to beast-users...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/beast-users/5ff56e0a-a38a-4510-89ff-b4ec546bfc76n%40googlegroups.com.

<不正确.jpg>