Is it preferable to start the analysis of my raw reads with Quality filtering first and after the paire-end assembly or the opposite ?
I’m not exactly sure what assembly you’re trying to do. But, in general, quality filtering and trimming should always be the first step in your analysis. Any assembly should definitely be done after filtering.
If you need any suggestions on a good program to do the quality filtering and trimming, I suggest Trimmomatic which very flexible and is already installed in the Compute Canada clusters.
Hi Genevieve, there is no universal answer as this is a question of where the best trade-off between sensitivity and specificity is for your particular target amplicon, biological sample, sequencing protocol, sequencing run quality, etc..
But in general some amount of prior quality trimming or hard clipping will help improve the assembling rate by removing high error regions.
I would recommend having a look at best practices, for instance through the dada2 tutorials: