I have 94 fasta files that I have merged into a single file with the merge.files command and analyzed with mothur. It returned 7000+ OTUs in total, averaging 400+ for each sample. That number of OTU seems weird to me, so I tried running one fasta file separately and got less than half of what merged fasta analysis coughed up:
sample A in merged fasta, -> 685 OTUs
Individual fasta, sample A -> 274 OTUs
I ran both analyses with the same commands, and the total seqs at the end were similar. So I wonder if merging fasta files inflating the number of OTUs is a norm?
PS: I'm using the latest version of mothur and their default settings outlined in MiSeq SOP.
i have not experience with mothur. But normally this kind of methods use the diferentiation with respect to other groups or the phylogenetic relationship. However, if your data are from different geographical region you can introduce some noice, specially for method based in differentiation, because is going to take as taxa similitude something that can be convergence. You did not give any detail of your data. Are they from different geographical regions?
Anyway, for that is recommended that you use different methods and depending of the nature of the data, you should separate or merge them. I will recommend to you more robust programs that already has been widely used for mOTU identification. The two more famous are:
ABGD: http://wwwabi.snv.jussieu.fr/public/abgd/
the nice of this program is that is showing you where are the gaps and you can also reconstruct a phylogenetic three. You can also play with the size of the gap in the settings.
GMYC (generalized mixed Yule-coalescent): http://species.h-its.org/gmyc/
is coalescence based and there are many other variation that you can try for your analysis.
then you can also take the data from mothu and compare results and see which one is the more conservative or logic.
I hope it was useful for you. Best,
Hi Patricia,
Sorry, I should mention my samples. I'm analyzing the gut bacteria of my crickets after giving them different diets, and it was conducted in a laboratory setting.
Thanks for recommending those programs. They are new to me and I will give them a try.
I have no experience in mothur but as per mothur tutorial, merge.files just concatenate fasta files in one file and do not process for any duplication. You can generate same file by "cat" command in terminal. Therefore, it is most probable that OTU_1, for example, is present in all x number of files, and merge.files will produce merged_file.fasta with x*OTU_1 times in sequence occurance.
And as per your final results, which considered the duplication of OTUs in merged file and took care of it, gave the same results for both types of analysis.
I hope this solved your confusion.
- Badges
- Science topic
- Similar topics
- Statistics
- Survey