I am new to NGS data analysis and I don't know how to overcome this problem: I have DNA-seq results from 12 different sites stored in one fasta file. Approximately every site has around 1000 reads, so I have 12000 reads in file but I don't know to which site any read belongs. I should build consensus sequences for all 12 sites without having reference sequence or knowing which reads belong to which site. Is it possible to do it and if yes, how?
An approach that doesn't require any prior knowledge is to cluster the sequences based on all pairwise sequence similarities/distances then compute a multiple sequence alignment for each cluster. Refinements can be made based on how much extra information is available.