Partitionfinder models for raxml

[Sithions Brand]

Hi evevyone,

RAxML allows for only a single model of rate heterogeneity in partitioned analyses. I.e. all partitions must be assigned one of three types of model: No heterogeneity (e.g. GTR); +G (e.g. GTR+G); or +I+G (e.g. GTR+I+G).    

If my mitochondrial sequence partitions each have different models, such as GTR+G, HKY+G, JC+G, and K80+I+G, how should I set the model parameters in RAxML? 

Best wishes

Sithions Brand 

[Oleksiy Kozlov]

Hi Sithions,

please use raxml-ng which supports individual model specification for each partition:

https://github.com/amkozlov/raxml-ng

https://github.com/amkozlov/raxml-ng/wiki/Input-data#multiple-models

Best,
Oleksiy

> --
> You received this message because you are subscribed to the Google Groups "raxml" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to
> raxml+un...@googlegroups.com <mailto:raxml+un...@googlegroups.com>.
> To view this discussion visit https://groups.google.com/d/msgid/raxml/7b5c1336-1bb1-4411-
> a8ae-456fe20ffafbn%40googlegroups.com <https://groups.google.com/d/msgid/raxml/7b5c1336-1bb1-4411-
> a8ae-456fe20ffafbn%40googlegroups.com?utm_medium=email&utm_source=footer>.

[Oleksiy Kozlov]

your partition file looks good.

please post FULL output of raxml-ng

On 13.11.24 07:59, Sithions Brand wrote:
> Dear Prof Kozlov,
>
> You're absolutely right about your suggestion, and I appreciate it. However, isn't it true that
> **RAxML-NG** doesn't support specifying codon partition models? Because in my **PartitionFinder**
> file, as shown in the image, I get an error during runtime saying the model cannot be found.
>
>
> my partitionfinder.txt:
> GTR+G, Subset1 = 1-976, 977-2568
> TrN+I+G, Subset2 = 12604-13965\3, 2569-3249\3
> TVM+I, Subset3 = 2570-3249\3, 12264-12603\3
> HKY+G, Subset4 = 2571-3249\3, 3250-3411\3
> TrN+I, Subset5 = 3251-3411\3
> GTR+G, Subset6 = 4965-5651\3, 5654-6435\3, 10248-11232\3, 3252-3411\3, 12606-13965\3, 14265-16071\3
> TIM1+I, Subset7 = 3412-4962\3
> TPM1uf, Subset8 = 3413-4962\3
> TrN, Subset9 = 3414-4962\3
>
> So I  get an ERROR: ERROR model initialization |partition.txt| (LIBPLL-5001): DNA model not found:
> partition.txt
>
> Best wishes
>
> Brand

> 在2024年11月13日星期三 UTC+8 10:08:22<Oleksiy Kozlov> 写道：
>
> Hi Sithions,
>
> please use raxml-ng which supports individual model specification for each partition:
>
> https://github.com/amkozlov/raxml-ng <https://github.com/amkozlov/raxml-ng>
>

> https://github.com/amkozlov/raxml-ng/wiki/Input-data#multiple-models <https://github.com/
> amkozlov/raxml-ng/wiki/Input-data#multiple-models>

>
> Best,
> Oleksiy
>
> On 12.11.24 15:54, Sithions Brand wrote:
> > Hi evevyone,
> >
> > RAxML allows for only a single model of rate heterogeneity in partitioned analyses. I.e. all
> > partitions must be assigned one of three types of model: No heterogeneity (e.g. GTR); +G (e.g.
> > GTR+G); or +I+G (e.g. GTR+I+G).
> >
> > If my mitochondrial sequence partitions each have different models, such as GTR+G, HKY+G,
> JC+G, and
> > K80+I+G, how should I set the model parameters in RAxML?
> >
> > Best wishes
> >
> > Sithions Brand
> >
> > --
> > You received this message because you are subscribed to the Google Groups "raxml" group.
> > To unsubscribe from this group and stop receiving emails from it, send an email to
> > raxml+un...@googlegroups.com <mailto:raxml+un...@googlegroups.com>.
> > To view this discussion visit https://groups.google.com/d/msgid/raxml/7b5c1336-1bb1-4411-

> <https://groups.google.com/d/msgid/raxml/7b5c1336-1bb1-4411->
> > a8ae-456fe20ffafbn%40googlegroups.com <http://40googlegroups.com> <https://groups.google.com/
> d/msgid/raxml/7b5c1336-1bb1-4411- <https://groups.google.com/d/msgid/raxml/7b5c1336-1bb1-4411->
> > a8ae-456fe20ffafbn%40googlegroups.com?utm_medium=email&utm_source=footer
> <http://40googlegroups.com?utm_medium=email&utm_source=footer>>.

>
> --
> You received this message because you are subscribed to the Google Groups "raxml" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to
> raxml+un...@googlegroups.com <mailto:raxml+un...@googlegroups.com>.

> To view this discussion visit https://groups.google.com/d/msgid/raxml/d27d4855-4220-436b-
> b880-6a273cc5ccbbn%40googlegroups.com <https://groups.google.com/d/msgid/raxml/d27d4855-4220-436b-
> b880-6a273cc5ccbbn%40googlegroups.com?utm_medium=email&utm_source=footer>.

[Sithions Brand]

Dear  Prof Kozlov,

Thank you for help!

Best wishes

Sithions