hello, everyone.
As I said, I have already finished my sequence amplification experiment of 4 markers of more than 200 samples. Now I have to calculate the species identification ability of each marker, which is the resolution of each markers as well. I am planing to do it with BLAST. As I understand, I need to build a database first, then use each sequence as a query to test the E-value of the blast ranked sequences and statistic analysis the resolution in the end.
I don't know if this is the right way to calculate, or is there any more detailed and correct instruction to tell me how to do with it.
By the way, the species is Quercus spp, a kind of plant.
Any suggestions will be appreciated! Thank you very much !
Hello Xiong,
There are two broad options for testing species identification using barcoding: 1) Percent Identity (PI) and 2) Genetic Distance (GD).
For PI most researchers use BLAST, however this method has been shown to either have similar or worst ability at correct identification and typically has far worst incorrect identification rate. Thus, using this method typically leads to a much higher probability of false identification then the GD methods. See Parmentier et al. (2011) for an example in rain forest trees (attached below). This method is also not robust to changes in the length of the sequences and is especially error prone with closely related species.
Thus, I would suggest using a GD method. These methods typically require building a global alignment and then gather distance data using a model of evolution. Then using these distances you can either build a phylogeny and look at what nodes sample fall in or just simply examining the matrix of distances and see what the closest match is. The Parmentier et al. paper has a good explanation of these methods and I also attached two other papers that are highly cited in the barcoding world that are great reads and may be helpful.
Best Regards.
Thank you very much, Max. It really confused me for a long time as I am new in bioinformatics, and your answer is really helpful !!!!
Thank you again, have a nice day !
HI Xiong,
another good idea is to construct phylogenetic trees with each of your markers. Ideally each of your species should be monophyletic (though there are exceptions!). You can also use automated methods like ABGD (http://wwwabi.snv.jussieu.fr/public/abgd/) which tries to identify and delimit species in your dataset based on genetic distances.
Best, Martin
- Badges
- Science topic
- Similar topics
- Biological Science
- Zoology
- Ornithology