Trimmomatic before or after or nope for genome guided assembly?

Question

[slin023]

Hi, I had experience doing De Novo assembly without trimming the paired-end libraries, but I am wondering what about trimmomatic applying to "genome guided assembly"?

If Before, you trim the libraries,remove the adaptors, mapping to the genome, create .bam files, then use Trinity to assembly transcriptome

If After you create mapping .bam files, then use Trinity to assembly transcriptome use --quality_trimming_params

or No ,not necessary, in case it over-trims the short reads?

I have 33 RNA seq paired-end libraries, 7 developmental time points, and 3 biological replicates for each 7 developmental time points

I am curious about what folks think about this trimmomatic idea on genome guided assembly, thank you for your time!

[GenoMax]

Any time you are doing de novo work you want to make sure that no extraneous sequence is in your data. Over-trimming may be ok (generally there should be plenty of data in your case here).