difference LG+C40 and LG+C40+G

[stra...@protist.eu]

Hi Bui Quang Minh,

Is there a difference between running -m LG+C40 and -m LG+C40+G? The manual says that C10 to C60 "assume Poisson AA replacement and implicitly include a Gamma rate heterogeneity among sites." Does that mean that gamma is employed automatically and there is no difference? If, however, there is a difference, why are the .rate files obtained with the following two commands exactly identical?

-s file.fasta -m LG+C40 -wsr -te allboottrees.suptree -n 0
-s file.fasta -m LG+C40+G8 -wsr -te allboottrees.suptree -n 0

I also realized that although G8 is called, only four categories are used?!

Thanks you!

[Minh Bui]

Hi there,

On 3 Jul 2018, at 2:11 am, stra...@protist.eu wrote:

Hi Bui Quang Minh,

Is there a difference between running -m LG+C40 and -m LG+C40+G? The manual says that C10 to C60 "assume Poisson AA replacement and implicitly include a Gamma rate heterogeneity among sites." Does that mean that gamma is employed automatically and there is no difference?

There is no difference between these two runs. LG+C40 implies LG+C40+G.

If, however, there is a difference, why are the .rate files obtained with the following two commands exactly identical?

-s file.fasta -m LG+C40 -wsr -te allboottrees.suptree -n 0
-s file.fasta -m LG+C40+G8 -wsr -te allboottrees.suptree -n 0

I also realized that although G8 is called, only four categories are used?!

Short answer: change to “-m LG+G8+C40” to have the wanted effect (8 rate categories).

Long explanation: If you run -m LG+C40+G8, you’ll see something like this in the log:

Model LG+C40+G8 is alias for LG+POISSON+G+FMIX{C40pi1:1:0.0223853788,C40pi2:1:0.0338891820,C40pi3:1:0.0577169375,C40pi4:1:0.0252416233,C40pi5:1:0.0108607921,C40pi6:1:0.0462373793,C40pi7:1:0.0102293175,C40pi8:1:0.0147523625,C40pi9:1:0.0143161352,C40pi10:1:0.0182302541,C40pi11:1:0.0204025079,C40pi12:1:0.0425505156,C40pi13:1:0.0248627269,C40pi14:1:0.0105892988,C40pi15:1:0.0188238725,C40pi16:1:0.0086663445,C40pi17:1:0.0148496147,C40pi18:1:0.0343037402,C40pi19:1:0.0225335203,C40pi20:1:0.0174068578,C40pi21:1:0.0112207827,C40pi22:1:0.0443532245,C40pi23:1:0.0392573370,C40pi24:1:0.0196756555,C40pi25:1:0.0287690328,C40pi26:1:0.0114441177,C40pi27:1:0.0112338740,C40pi28:1:0.0582694099,C40pi29:1:0.0444272279,C40pi30:1:0.0112010942,C40pi31:1:0.0145176111,C40pi32:1:0.0114629026,C40pi33:1:0.0239628061,C40pi34:1:0.0266266492,C40pi35:1:0.0481201159,C40pi36:1:0.0371147423,C40pi37:1:0.0160476688,C40pi38:1:0.0237249267,C40pi39:1:0.0235226203,C40pi40:1:0.0261998398}+G8

While parsing this full model string from left to right, “+G” occurs first before “+G8”, and therefore it is taken as Gamma model with default 4 categories.

If you now do LG+G8+C40, then you’ll see:

Model LG+G8+C40 is alias for LG+G8+POISSON+G+FMIX{C40pi1:1:0.0223853788,C40pi2:1:0.0338891820,C40pi3:1:0.0577169375,C40pi4:1:0.0252416233,C40pi5:1:0.0108607921,C40pi6:1:0.0462373793,C40pi7:1:0.0102293175,C40pi8:1:0.0147523625,C40pi9:1:0.0143161352,C40pi10:1:0.0182302541,C40pi11:1:0.0204025079,C40pi12:1:0.0425505156,C40pi13:1:0.0248627269,C40pi14:1:0.0105892988,C40pi15:1:0.0188238725,C40pi16:1:0.0086663445,C40pi17:1:0.0148496147,C40pi18:1:0.0343037402,C40pi19:1:0.0225335203,C40pi20:1:0.0174068578,C40pi21:1:0.0112207827,C40pi22:1:0.0443532245,C40pi23:1:0.0392573370,C40pi24:1:0.0196756555,C40pi25:1:0.0287690328,C40pi26:1:0.0114441177,C40pi27:1:0.0112338740,C40pi28:1:0.0582694099,C40pi29:1:0.0444272279,C40pi30:1:0.0112010942,C40pi31:1:0.0145176111,C40pi32:1:0.0114629026,C40pi33:1:0.0239628061,C40pi34:1:0.0266266492,C40pi35:1:0.0481201159,C40pi36:1:0.0371147423,C40pi37:1:0.0160476688,C40pi38:1:0.0237249267,C40pi39:1:0.0235226203,C40pi40:1:0.0261998398}

Since +G8 appears first in this model string, it “tricks” IQ-TREE to interpret it as 8 Gamma categories.

Hope that helps,

Minh

Thanks you!

--
You received this message because you are subscribed to the Google Groups "IQ-TREE" group.
To unsubscribe from this group and stop receiving emails from it, send an email to iqtree+un...@googlegroups.com.
To post to this group, send email to iqt...@googlegroups.com.
Visit this group at https://groups.google.com/group/iqtree.
For more options, visit https://groups.google.com/d/optout.

[stra...@protist.eu]

Hi,

Thanks for the fast reply!

So it is impossible then to run iq-tree with LG+C40 only? I am asking since the model test favored  LG+C40 over LG+C40+G. What would iq-tree do if I run a tree with -m TEST (and the best model is C40 without gamma)? Would it ignore its own test and run it with gamma?

Best, Jurgen

[stra...@protist.eu]

I am attaching the model test here to make clear what I mean. How should I interpret the output when C40 and C40+G is the same and why are the values different then?

[Minh Bui]

Hi Jurgen,

Can you send me the log file via my personal email? Will have closer look to see why this output…

Minh

<Output_telo.fasta.iqtree>

[stra...@protist.eu]

Hi Minh,

I have one more question that is related to the one below. I understand now that gamma with default 4 categories is used when it is not specified at first appearance. But what happens with +F? Let's say I used:

-s file.fasta -m LG+C40+G8+F -wsr -te allboottrees.suptree -n 0

Will +F still be considered or be ignored just like +G8 because it is at a wrong position in the string?

Thank you!

/J

Am Dienstag, 3. Juli 2018 05:07:16 UTC+2 schrieb Minh Bui:

[Minh Bui]

Hi Juergen, +F will be considered. I.e. your model below is a profile mixture model with 41 classes, 40 from C40 and one from +F class. The discussion was only concerning +G.

Cheers

Minh