search

ansariazim / treeBreaker

Algorithm to divide a phylogenetic tree into segments based on phenotypes at the leaves of the tree
25 stars 3 forks source link

Requirements for the input tree? #6

Closed maesaar closed 3 years ago

maesaar commented 4 years ago

Is there some requirements for the input tree?

FIRST: Is it important to use whole-genome alignment or is core genome alignment also suitable for tree generation?

SECOND: What tree would be best to use as an input: 1) wgMLST core genome tree (with recombination)? 2) The same tree but with corrected recombination eg ClonalFrameML output tree? 3) Output tree from BactDating using CFML tree as input?

Thank you for the clarifications

maesaar commented 4 years ago

As I understand from the manuscript unrecombined core genome is suitable to use. But question still remains is it ok to use CFML output tree or BactDating output tree?

The manuscript said phenotypic trait was used, but would it also work with genotypic trait (coded 1,0)?

Thank you again

ansariazim commented 4 years ago

Hi there,

FIRST: The method does not depend on how you have generated the tree, it works on any kind of tree. What kind of tree you should use and does it make sense depends on the questions that you are asking.

SECOND: Again this depends on the question that you are asking. The method does not assume anything about how the tree was generated. It assumes that the tree is the correct tree for your question. So you could potentially use all of those trees, You need to think about what tree answers the questions that you are asking.

Best. Azim

On 16 Jun 2020, at 22:20, mmaesaar notifications@github.com wrote:

Is there some requirements for the input tree?

FIRST: Is it important to use whole-genome alignment or is core genome alignment also suitable for tree generation?

SECOND: What tree would be best to use as an input:

wgMLST core genome tree (with recombination)? The same tree but with corrected recombination eg ClonalFrameML output tree? Output tree from BactDating using CFML tree as input? Thank you for the clarifications

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/ansariazim/treeBreaker/issues/6, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAQDHEGIPL3KCBJL2ORHXDTRW7OZFANCNFSM4N77YTFQ.

ansariazim commented 4 years ago

Hi there,

Yes, you can use genotypic trait or any other trait. It will find branches where the trait distribution has changed relative to the rest of the tree.

Best. Azim

P.S. If you want to use treeBreaker, please use the version with the last commits from 2016. the R version does not seem to install properly for people. So you need to use the command line version from 2016.

On 16 Jun 2020, at 22:33, mmaesaar notifications@github.com wrote:

As I understand from the manuscript unrecombined core genome is suitable to use. But question still remains is it ok to use CFML output tree or BactDating output tree?

The manuscript said phenotypic trait was used, but would it also work with genotypic trait (coded 1,0)?

Thank you again

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/ansariazim/treeBreaker/issues/6#issuecomment-645023449, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAQDHEHZQRAVZYRE7YFWW53RW7QKBANCNFSM4N77YTFQ.

maesaar commented 4 years ago

Thank you for thorough answer.

I am currently trying to use treeBreaker using virulence and amr genes with dated unrecombined core genome tree to see when did these genes emerge.

Would it be better if genotypic trait (eg gene/mutation) is present in alignment what would be used for tree generation or not?

Thanks

ansariazim commented 4 years ago

Hi there,

I would say if you are using core genome, then including a gene or two in your alignment is not going to change the tree much at all. So it probably is not going to make any difference if you include them or not in practice.

In principle if the tree is supposed to capture the clonal genealogy between the isolates, then the question would be do you think including the gene that you are interested in, improve or weaken the clonal genealogy inference?

Sorry, I cannot give you a definite answer, as I am not familiar with your data. But in practice I would say it is not going to make that much difference if you include them or not.

Best. Azim

On 17 Jun 2020, at 11:43, mmaesaar notifications@github.com wrote:

Thank you for thorough answer.

I am currently trying to use treeBreaker using virulence and amr genes with dated unrecombined core genome tree to see when did these genes emerge.

Would it be better if genotypic trait (eg gene/mutation) is present in alignment what would be used for tree generation or not?

Thanks

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/ansariazim/treeBreaker/issues/6#issuecomment-645298258, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAQDHEBDSVQZU4SUMXVQP2LRXCM5ZANCNFSM4N77YTFQ.