Dear Colleagues,
I need some advice to resolve the issues associated with building robust phylogeny trees in different fungal families. It is indeed a basic query but may be helpful to many budding mycologists from India and across the globe.
We are using graphic version of RAxML and MrBayes for making trees.
Unfortunately, we are regularly facing issues related to data deficiency in several fungal families ( It cannot be resolve until sampling and epitypification!!!).
But even after retrieval of data based on some published papers; The BS/BYPP values seem to vary considerably from the high values mentioned in the article to a moderate value in our analyses.
What are the best practices to deal with it?
We choose type strain to build a backbone to assess the topological congruence for different genes too.
How best we must edit our datasets to achieve high BS values? How best we can define gaps in RAxML?
What are the other parameters taken into account other than model tests?
Thanks in advance.
Yeah, a nice question.
I am a PhD student majoring in Bioinformatics. I have similar problems with Rajeshkumar K. C. . Sometimes even I used same dataset, I couldn't get the similar phylogenetic tree， even with partially incongruent topology.
I think that the difference in alignment tools, tree inferring tools, evolution models, the order of linking multiple markers and trimming methods maybe lead to the incongruence. Specially, I realized that many authors used "manually trimmed" to describe the trimming process, but we do not know what sites were removed. Every one has his own idea in trimming alignments.
In my opinion, no mater what software you use, the author should confirm that, other people can reproduce the phylogenetic analysis based on the original data (sequence accession table) not the alignment files, according to the method description.
In practice， I prefer to use trimal to trim the alignment files, and other people can get identical alignment files according my command options. We can confirm that other people can get most similar or even identical results according to our method description.
One of the major reason is manual data set editing that leads to differences in bootstrap values as well as topologies.
You can access to IQ-TREE website, it maybe a good option, here is the link. http://iqtree.cibiv.univie.ac.at/
A phylogeny is predicted model of an evolutionary history of organisms (for molecular phylogeny, it is history of gene or genes) based on observation data from current species. Some of these histories simply can not be robustly resolved with current data, and there is not much you can do. Otherwise, to chose the right data for right time (relationship level) is better than just to combine all data available. Please refer to "Phyloinformativeness" approaches developed by Dr. Townsend (Article PhyDesign: An online application for profiling phylogenetic ...
).- Badges
- Science topic
- Similar topics
- Phytochemistry
- Extracts