Dear Biostar Community,
Currently I'm doing a phylogenetic study on 37 samples.
The samples were sequenced using MIG-seq.
The sequencing data is in fastq format.
Using Julien Catchen's Stacks, I was able to create a SNP matrix. From the SNP matrix, I was able to do a PCA and find that there are three tightly grouped clusters. Additionally, the tree built by RAXML has three major clades all with bootstrap values of 100.
Thus, I can say with relative certainty that I'm dealing with three separate species. Based on some morphological data, I also have a fairly good guess on what those species are. I'd like to confirm the identity of these species genetically.
My plan of attack was to download (from NCBI) various genomes belonging to the sample's genus, build a SNP matrix with these downloaded genomes and my 37 samples, and rerun PCA and RaxML with the expectation that the proper genome will be clustered with samples of the same species.
I'm struggling with the right way to do this though.
I tried to use Stacks (denovo) to build a catalogue of loci for my 37 samples. I then used bowtie2 to align my downloaded genomes to that catalogue of loci with the hope that this will create a usable .bam output that I can then use in Stacks to compare against the 37 samples. The problem is that no matter how I tweak bowtie2's parameter's, the run will inevitably eat up all the computer's memory and eventually throw an error.
Currently, I'm using RADinitio to simulate sequencing the downloaded genomes via RAD-seq and then use those fastq outputs to run with my 37 samples. However, this seems like a backwards and roundabout way of doing things.
It feels like there should be a simple enough solution to this question, but I can't seem to find anything online. Any help or direction would be greatly appreciated. Thank you!
