I have RNA-seq data from 4 experimental conditions, each one with 3 replicates, and my primary goal is differential expression. My doubt is if combining all the samples into a single input target for the transcriptome assembly would be recommended for all assemblers.
I was using Trinity and they do suggest to combine all samples into a single fastq file for performing the analysis. However, I also want to test other methods (such as SOAPdenovo-trans, Oases). Unfortunately I couldn't find clear information if merging the fastq files would also be recommended for them. If not, why?