I am an evolutionary biologist with a broad interest in computational and statistical methods for understanding evolution of organisms and patterns of biodiversity.

-Statistical methods for species delimitation

One of my primary research interests is developing methods for finding species from DNA sequences, that is, DNA-based species delimitation. Not only it is simply cool if you can find new species only from DNA sequences, but it can provide researchers with a rapid and universal way to quantify unknown biodiversity, which is useful under the current condition of rapid biodiversity loss.

The Generalized Mixed Yule Coalescent (GMYC) model is one of the most widely used DNA-based delimitation methods for single locus sequence data such as DNA barcoding. I worked on the GMYC method, originally developed by my supervisor, Tim Barraclough, in my PhD study. I added some computational extensions to the original version and  implemented its R package, “splits”.


An example of delimitation with GMYC. Threshold time is indicated by a red line in the left panel. Delimited species are colored in red in the right panel.

In a more recent work, I developed a multilocus delimitation method, which can leverage modern high-throughput sequence technologies. The Trinominal distribution of Triplets (Tr2) model can delimit species with sequences of hundreds of loci, which allows you to delimit more recently diverged species than conventional single locus methods.


Multilocus delimitation of Bacillus isolates using tr2-delimitation. (Fujisawa et al. 2016)

As more and more sequence data are becoming available, we are in need of finer scale delimitation methods which incorporate more realistic statistical models of demography of populations and heterogeneous genome evolution. I am now aiming at developing such methods.

Selected publications

Fujisawa and Barraclough (2013) Delimiting species using single-locus data and the Generalized Mixed Yule Coalescent (GMYC) approach: A revised method and evaluation on simulated data sets. Syst.Biol. 62(5): 707-724. 

Fujisawa, Aswad and Barraclough (2016) A rapid and scalable method for multilocus species delimitation using Bayesian model  comparison and rooted triplets. Syst.Biol. 65(5): 759-771.

-Speciation genomics

Development of species delimitation methods is closely related to studies of speciation process itself. How discrete units of organisms evolve has long been an important research topic for evolutionary biologists. Researchers have been studying key traits which drive species divergence, such as life cycle, body size, genital morphology and so on.  I approach this topic with data analysis of  modern high-throughput sequence technologies and statistical methods, which give us ways to reveal detailed histories of species divergence and underlying genome evolution.

I worked in Teiji Sota’s group in Kyoto university, Japan and analysed multiple data of high-throughput sequence technologies including transcriptome, RAD and whole genome sequence. Targets include the genes responsible for controlling body size of beetles and life cycles of cicadas.


Quantitative trait locus analysis combined with whole genome sequencing revealed functional groups of genes involved in the body size divergence of Carabus beetles. (Komurai et al. 2017)

Selected publications

Komurai, Fujisawa, Okuzaki and Sota (2017). Genomic regions and genes related to inter-population differences in body size in the ground beetle Carabus japonicus. Scientific Reports, 7(1), 7773.

Other manuscripts are under review and in preparation…

-Molecular evolution of mitochondria and sequence databases

Public databases of DNA sequences and biodiversity information are ever-growing. I am interested in using the global standardized data deposited in these public databases to ask questions of molecular evolution.

My works have mainly focused on the evolution of mitochondrial genes, which is the most widely used molecular markers. Data mining of sampling records and geo-referenced sequences enabled a comparative analysis of correlates of molecular evolution at large scale. In a paper, I investigated the effect of the long-term patterns of evolution of habitat types affect the speed of molecular evolution using a barcoding-type data set of beetles.


Bayesian phylogeny of GMYC-delimited species of diving beetles with their habitat types (Fujisawa et al. 2015)



Effect of habitat type (lotic vs. lentic) over the speed of COI evolution. We found the lotic species evolve faster after correcting the effect of node density. (Fujisawa et al. 2015)

Also, I worked with Alfried Vogler in NHM, UK and Andres Baselga in USC, Spain to investigate a unified mechanism controlling geographic distributions of species and haplotypes, which is a more ambitious project to bridge community ecology and community genetics.


Distance decay of assemblage similarity of different genetic hierarchical levels. (Baselga et al. 2013)

Now, the amount of data in public databases are rapidly increasing and available data include not just DNA sequences but more general biodiversity information (such as sampling records or specimen photographs). I am interested in using these data sets and advanced statistical techniques to extend this type of approach.


Fujisawa, Vogler and Barraclough (2015) Ecology has contrasting effects on genetic variation within species versus rates of molecular evolution across species in water beetles. Proc. R. Soc. B. 282:20142476

Baselga, Fujisawa, Crampton-Plat, Bergsten, Foster, Monaghan and Vogler (2013) Whole-community DNA barcoding reveals a spatiotemporal continuum of biodiversity at species and genetic levels. Nat. Commun. 4:1892