Combining morphological characters and molecular characters into the same phylogenetic analysis can be done in MrBayes (if you are ok with a Bayesian approach), regardless of missing taxa or characters.

You make one matrix including both types of data. I make a nexus file in a text editor like Text Wrangler. The matrix is then partitioned to assign the different rate models to the different kinds of data (the same way you would for a multigene dataset). Here is an example of a MrBayes block I wrote for my matrix of 7 genes and a set of morphological characters:

begin mrbayes;

    charset 12S = 1-376;

    charset 16S = 377-987;

    charset cytb = 988-1701;

    charset RAG1 = 1702-3221;

    charset NCX1 = 3222-4532;

    charset SLC8A3 = 4533-5693;

    charset CXCR4 = 5694-6332;

    charset morphology = 6333-6444;

    partition favored = 8: 12S, 16S, cytb, RAG1, NCX1, SLC8A3, CXCR4, morphology;

    set partition = favored;

    outgroup AscaphusT;

    lset applyto=(1,2,3,4,6,7) nst=6 rates=invgamma;

    lset applyto=(5) nst=2 rates=invgamma;

    lset applyto=(8) rates=gamma;

    unlink statefreq=(all) revmat=(all) shape=(all) pinvar=(all);

    prset applyto=(all) ratepr=variable;

    mcmcp ngen=5000000 samplefreq=1000 printfreq=1000 savebrlens=yes;

end;

The MrBayes wiki page describes this in detail as well and will explain some of the lines of code if these are unfamiliar to you. Hope this helps! 

I am not sure but hope PAUP will give you better results and will combine two sets. GenAlex will give you PCA using Morphological and Molecular data set. Try these

Thanks Dr. Padder. Yes, PAUP is a good tool to analyze combined data but what I need is a strategy to combine the two types of data. Do you have any opinion on this?

Hi

What about having two separate dendrograms with similar software (for instance; Genealex, Paup, Mega5, NTYsis and etc and compared two distinct ones?

Thanks Atena,

I've prepared two trees using the two data sets, however, my concern is that the trees differ from each other due to the presence of additional taxa. Further, the molecular data gives a tree congruent to the tree explained in other studies on the family from other regions but the tree from the morphological characteristics doesn't.

I thought combining morphology and molecular data would address this to a certain extent.

if you want to combine two data sets then try NTYSIS. Develop two matrices and using MAXCOMP option to combine the both. The other way is to use missing data for samples for which DNA profile is not with you. By this way you should be able to combine both data sets.

I believe a partitioned analysis in MrBayes will do this for you.

Edit: You may need to separately derive the tree based on morphological data only, and then integrate that into the tree generated from a partitioned analysis of both morphological and genetic data, but check out the manual.

your welcome

I think this is natural, because the changeability of morphological and molecular characteristics have different speed. Most of the time 2 species separated morphologically but not molecularly. morphological properties are adujested more quickly than the other one.

Thanks Dr. Moore. I've tried using MrBayes for this, however, I could not figure out a way to combine the trees (or data). Your explanation provides some insight to it. I will check out the manual too.

Dr. Padder, I'm not familiar with NTYSIS, may be I'll take a look at it too. thanks though.

Thanks Dr. Guerin. I didn't use R for my analysis though I'm familiar with the techniques in phylogenetic analysis with R.

I'll give it a try.

Regards.

I remeber that the the study of placental mammal evolution by O'Leary et al. (2013) combined molecular and morphological data. The paper was published in Science 339 and given the space confinements in Science they do not give much detail about the contruction of the matrices but there should be something about their approach to combine molecular and morphological data in the supplemantary matrial (which I haven't read yet).:

Here is the link top the supplementary matrial for the paper:

www.sciencemag.org/cgi/content/full/339/6120/662/DC1

I hope this helps you along.

As far as I understand your problem, this is primarily an issue of missing data. The combination of two different types of data is feasible, as long as you are able to manage the balance of importance of these two (or more) subsets (weighing). See also the answer of Jonathan Moore.

if you have two types of data (A and B) and a part of your species P show both data A and B, and another set of species Q have only data A, then your full data matrix shows four quadrants: P-A, P-B, Q-A, and the empty Q-B. Run two analyses: one for data A with all species P+Q, and one for data B with only species P. Compare tree B with the part of tree A where the species P are located. If you have also lacking data for set A and, hence, a set of species R with lacking data A, then your data matrix has six quadrants and two empty ones (R-A and Q-B). Then a full analysis is virtually impossible: the structure of your tree is largely directed by your missing data and their weighing.

However, there is a solution in your case. If you can collect morphological data from those herbarium specimen of presumed extinct taxa, then it is also possible to collect DNA. These techniques are nowadays fully developed, and this allows you to make your data matrix complete, or at least less incomplete than it is now.

Combining morphological characters and molecular characters into the same phylogenetic analysis can be done in MrBayes (if you are ok with a Bayesian approach), regardless of missing taxa or characters.

You make one matrix including both types of data. I make a nexus file in a text editor like Text Wrangler. The matrix is then partitioned to assign the different rate models to the different kinds of data (the same way you would for a multigene dataset). Here is an example of a MrBayes block I wrote for my matrix of 7 genes and a set of morphological characters:

begin mrbayes;

    charset 12S = 1-376;

    charset 16S = 377-987;

    charset cytb = 988-1701;

    charset RAG1 = 1702-3221;

    charset NCX1 = 3222-4532;

    charset SLC8A3 = 4533-5693;

    charset CXCR4 = 5694-6332;

    charset morphology = 6333-6444;

    partition favored = 8: 12S, 16S, cytb, RAG1, NCX1, SLC8A3, CXCR4, morphology;

    set partition = favored;

    outgroup AscaphusT;

    lset applyto=(1,2,3,4,6,7) nst=6 rates=invgamma;

    lset applyto=(5) nst=2 rates=invgamma;

    lset applyto=(8) rates=gamma;

    unlink statefreq=(all) revmat=(all) shape=(all) pinvar=(all);

    prset applyto=(all) ratepr=variable;

    mcmcp ngen=5000000 samplefreq=1000 printfreq=1000 savebrlens=yes;

end;

The MrBayes wiki page describes this in detail as well and will explain some of the lines of code if these are unfamiliar to you. Hope this helps! 

Further to Hilary's answer, when combining data in MrBayes you also need to change the Datatype at the top of the nexus file. Here is an example from one of my analyses:

BEGIN DATA;

    DIMENSIONS NTAX=78 NCHAR=25785;

    FORMAT DATATYPE = MIXED (DNA:1-25515, STANDARD: 25516-25785) INTERLEAVE GAP = - MISSING = ?;

Here I've also used an interleaved format, where the morphological charcters are in a separate block to the DNA. This just makes things easy to inspect. I believe there should be a nexus file that clarifies how to set things up associated with one of my supervisor's papers - a phylogeny of golden moles (Asher et al. 2010 - A phylogenetic estimate for golden moles (Mammalia, Afrotheria, Chrysochloridae, BMC Evol. Biol.). The DOI is 10.1186/1471-2148-10-69, and its open access, so you should able to see it.

In terms of your data itself, for the combined analysis approach to work you will need morphological characters for the majority of taxa, not just those that lack DNA (I'm sure you know this). This will hopefully result in the morphological character optimisation being influenced by the DNA in those taxa that have both character sets, placing your morphology-only taxa in a phylogenetic position influenced by your genetic dataset.

I hope that is in someway helpful!

If you want to integrate morphological and molecular data into tree building I would recommend this paper: http://sysbio.oxfordjournals.org/content/early/2014/04/28/sysbio.syu022.abstract

About DNA extraction from old herbarium sheets I recommend this paper DNA Extraction from Herbarium Specimens by Lenka Záveská Drábková https://www.researchgate.net/publication/259700673_DNA_extraction_from_herbarium_specimens

Article DNA Extraction from Herbarium Specimens

You can also make two different trees, using the methods os your choice for each dataset and then combine trees with supertree approaches. MRP is the most common, but you should also try sfit (on clann program). sfit is super fast and I believe it works even better than MRP.

POY has also a very interesting way to combine molecular and morphological data.

http://www.academia.edu/2685025/POY_5.0_RC

https://code.google.com/p/poy/

Be careful here.  I think Leo is correct.  The problem here is missing data.  You do not have sequence data for those taxa so you are relying just on the morphological characters for placement.  I would try the analysis of the combined data set but just for comparison also generate trees from the molecular data set and the morphological data set separately.  This may help illuminate the information content of the morphological characters.

Did somebody use the Mesquite for combining molecular and morphological data? Do you know a algorithm  for doing that?

Hello everyone,

I don't know if you only want to do the search using Bayesian Inferences but if you don't, I think your best option is to combine the data and perform the analysis with TNT (by parsimony). It is a very friendly and fast software (non command-line, but menu interface). Best regards!.

I added a Nexus file with mixed data, made in Notepad:

I found this short handout from a course taught at UC Berkeley very helpful: http://ib.berkeley.edu/courses/ib200a/ib200a_sp2008/ConcatenatingDataSets.pdf . I have also uploaded a copy in case the link becomes invalid.

Moleculaer data reveal a pure phylogeny.

The morphology is a product of convergent evolution.

These two beings are principially incongruent.

Their summation leads to unconcretization of taxa evolutionary status.

Yes indeed