RaxML-NG, bootstrap, Cipres

[Ari]

Hello again.

I want to perform a standard non-parametric bootstrap with 1000 replicates (without bootstopping) in RaxML-NG. Given I cannot set the replicates number, nor turn off the bootstoping option, I suspect that the only option in Cipres is run a maximum of 1000 replicates plus bootstopping (autoMRE, cutoff: 0.030000). That is correct?

Thank you

Best regards

[Wayne Pfeiffer]

Hi Ari,

Yes, the only bootstrap option currently exposed via CIPRES seems to be the default one: bootstopping with autoMRE and a maximum of 1000 bootstraps. I will follow up with Mark about allowing a specific number of bootstraps to be specified.

Incidentally, bootstrap searches are much slower in RAxML-NG than in standard RAxML. If that is of concern for you, standard RAxML might be a better choice for the bootstraps. If you mix the code versions in your analysis, you should note that RAxML-NG uses a ML estimate for the base frequencies (F0) by default, whereas standard RAxML uses empirical frequencies (F or FC) by default.

Best regards,  Wayne

--
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.
To view this discussion on the web visit https://groups.google.com/d/msgid/raxml/457c5778-abcc-4663-ac2c-9198903bd042o%40googlegroups.com.

[Ari]

Thank you for your response!

Best regards

El domingo, 20 de septiembre de 2020, 23:49:03 (UTC-3), Wayne Pfeiffer escribió:

Hi Ari,

Yes, the only bootstrap option currently exposed via CIPRES seems to be the default one: bootstopping with autoMRE and a maximum of 1000 bootstraps. I will follow up with Mark about allowing a specific number of bootstraps to be specified.

Incidentally, bootstrap searches are much slower in RAxML-NG than in standard RAxML. If that is of concern for you, standard RAxML might be a better choice for the bootstraps. If you mix the code versions in your analysis, you should note that RAxML-NG uses a ML estimate for the base frequencies (F0) by default, whereas standard RAxML uses empirical frequencies (F or FC) by default.

Best regards,  Wayne

On Sep 19, 2020, at 11:40 PM, Ari <ari...@gmail.com> wrote:

Hello again.

I want to perform a standard non-parametric bootstrap with 1000 replicates (without bootstopping) in RaxML-NG. Given I cannot set the replicates number, nor turn off the bootstoping option, I suspect that the only option in Cipres is run a maximum of 1000 replicates plus bootstopping (autoMRE, cutoff: 0.030000). That is correct?

Thank you

Best regards

--
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 ra...@googlegroups.com.

[Ari]

Hi.

It seems that the bootstoping option is not currently available for "--all" analyses. I am noticed that you are right: some bootstrap searches take too much time. Is there any possibility to return the previous default options?

Sorry for the inconvenience!!

Thank you

[Sara Rocha]

Hi,

related question, is there a way in RaxML-NG to have the boots written to a file and not only in the end? same for ML tree (it would help to have a look at say 100, when you actually ask for 1000 given that they take so long...)

best, and thanks in advance...

[Alexey Kozlov]

Hi Sara,

the latest version will do it by default, so please upgrade ;)

https://github.com/amkozlov/raxml-ng/releases/tag/1.0.1

Best,
Alexey

>> <https://urldefense.com/v3/__https://groups.google.com/d/msgid/raxml/457c5778-abcc-4663-ac2c-9198903bd042o*40googlegroups.com?utm_medium=email&utm_source=footer__;JQ!!Mih3wA!XBDerBTwXrLyDBwy_KfI5oPDlPItrJ1WUYfnLgblgAdEw6lTiPLjBfkJbxUJ96wQ$>.

>
> --
> 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 on the web visit

> https://groups.google.com/d/msgid/raxml/ef453df4-061d-4054-afe3-f3f03b312f17n%40googlegroups.com
> <https://groups.google.com/d/msgid/raxml/ef453df4-061d-4054-afe3-f3f03b312f17n%40googlegroups.com?utm_medium=email&utm_source=footer>.