Dear Martin,

You can use the R package 'abc' from Csillery et al (http://onlinelibrary.wiley.com/doi/10.1111/j.2041-210X.2011.00179.x/abstract). You can easily find in the internet an R vignette explaining with examples the different functions of the package through examples (i.e., different methods of parameter estimation/model comparison, posterior predictive checking, cross-validation procedures...). The only thing that you cannot do with this package is to simulate the microsatellite data, but you can use other command line programs for simulate the data (e.g. simcoal, ms) and make the abc analysis with this package.

This is what we have done in the following papers for instance (http://onlinelibrary.wiley.com/doi/10.1111/mec.13401/full and http://onlinelibrary.wiley.com/doi/10.1111/mec.13345/full), where we used the flexible ABCtoolbox framework of Wegmann et al 2010 (BMC Bioinformatics, 11, 116) to build a pipeline of different programs (i.e. simcoal, PDGSpider, arlequin, ADZE and R) that allowed us to calculate many different summary statistics (including the M-ratio/GW statistic) for our simulations. Then, we analysed the output with the abc package.

A similar approach was also used by Marino et al 2013 (http://onlinelibrary.wiley.com/doi/10.1111/mec.12458/full).

You can also call the ms program of Hudson et al 2012 (Bioinformatics, 18, 337–338) as an external program with R, then analyse the output with the package 'abc' or others. This was done in http://onlinelibrary.wiley.com/doi/10.1111/mec.12321/full, and an R script for launching ms from R is available at the dryad entry associated to this last article.

Hoping that you find this useful,

Ivan

Hi Ivan,

thanks for your excellent answer! This helps a lot!

cheers,

Martin

Hello Martin,

Some  collaborators and I gave a course about model selection in the last Evolution Meeting. In this website you can found all the resources we used for this course,  some scripts to perform ABC with microsatellites and with sequence data: http://2015nsfssbworkshop.weebly.com/resources.html. The lectures were videotaped, and you can find a link to the videos in this same website.

I also published a recent paper using ABC in two plant species: http://www.nature.com/hdy/journal/v116/n6/full/hdy201617a.html. I made all the scripts that I used available in dryad: http://datadryad.org/resource/doi:10.5061/dryad.8h6k1.

If you need any help with these scripts you can let me know. All the analyzes were based on simulation with the software 'ms', summary statistics were calculated with 'Arlequin' and the ABC itself was performed with Csillyry's et al package 'abc' mentioned by Dr. Paz-Vinas in the previous message.

Hope this can help you.

Manolo, this is perfect, I have a look at these resources, thanks alot!

Manolo Perez the link for the 2015 course you posted is now broken - do you have an updated link? Thank you!

Dear Maria Dance ,

I uploaded the contents of the course in my personal website. You can access it here: https://sites.google.com/site/manolofperez/talkslectures.

Please let me know if you have any problems with these scripts.

Best,

Manolo Perez Thank you very much for the quick reply but this link doesn't seem to work either (see picture)

Maria Dance , that's strange because it's working on my computer. I'll check what's the problem.

try this link for the dataset itself: https://1drv.ms/f/s!An30RV_2QfTigvcRcQOvPopDEtxgpQ.

Thank you, and sorry for that.

Thank you so much Manolo Perez - this is great. I'm new to abc so this will be really helpful!

Thanks Manolo Perez, I am also new to ABC and I need it for one of my PhD chapters. These Videos and notes will be very helpful.

There is a paper on arxiv now for this: Preprint A Guide to General-Purpose Approximate Bayesian Computation Software

Good luck!

Manolo Perez . Hi I hope you are doing good. I am looking at ABC Model selection. I fallowed the most common tutorial/example to do model selection. Based on results it says, These results show that Hausa data supports best the model of exponential growth model, while the Italian and Chinoise data supports best the population bottleneck model. I really can't figure it out that how to conclude best model based on matrix table I get after executing summary function. There is no further explanation of it either. In my case I have I want to do model selection of three different models.

Call:

postpr(target = voight["1", ], index = mod3, sumstat = pop, tol = 0.8,

method = "mnlogistic")

Data:

postpr.out$values (432 posterior samples)

Models a priori:

LHSMDU, LHSPSO, PSO

Models a posteriori:

LHSMDU, LHSPSO, PSO

Proportion of accepted simulations (rejection):

LHSMDU LHSPSO PSO

0.4167 0.1667 0.4167

Bayes factors:

LHSMDU LHSPSO PSO

LHSMDU 1.0 2.5 1.0

LHSPSO 0.4 1.0 0.4

PSO 1.0 2.5 1.0

Posterior model probabilities (mnlogistic):

LHSMDU LHSPSO PSO

0.1464 0.0000 0.8536

Bayes factors:

LHSMDU LHSPSO PSO

LHSMDU 1.000000e+00 3.299767e+08 1.715000e-01

LHSPSO 0.000000e+00 1.000000e+00 0.000000e+00

PSO 5.830300e+00 1.923866e+09 1.000000e+00

If you can help me out in above result. thanks

Hi, lookf or the abc package for R, by Csillery. The package has a nice manual, and the authors has published papers on it. Also, the paclkage does not make the simulations of summary stats, which you would need to obtain elsewhere.

Best

Dear Muhammad Usman,

Based on the output you presented, the best model would be PSO when you use mnlogistic (PP = 0.8536), while the regular rejection was only able to point that LHSPSO was the worst model, as the other two models had the same PP (0.4167).

Note, however that there are some issues with your analysis. Your threshold (tol = 0.8) is way too high in my opinion. And I would say that your number of simulations is too small for an ABC analysis, at least in my opinion. I suggest you to try a number of simulations and parameters in the postpr command more similar to the ones in the tutorial you mentioned, that's a good start.

Hope this can be helpful.

Here is the updated link for the ABC course: https://1drv.ms/f/s!An30RV_2QfTigvcRcT6AvB2r0SW2VQ.