Illumina Reads Quality Control and Preprocessing

[Lisa0508]

Hi,every warm-hearted people in the forum. I am a beginner on genome assembly. I have been using Fastx-toolkit to remove the reads that have less than 80% of bases with quality lower than a value of 20, and then mask the bases below 20 with 'N'. But I did not the trimming function because I'm thinking about cleaving a read at 'N' site, removing 'N's and spliting the original read to several fragments so that I can keep more reads. Finally, I will eliminate the fragments below eg,50bp. But thus the header(ID name) of those fragments will be a problem. I think if I just use those reads for de novo assembly, I don't need to split at 'N' because the reads containing N will be k-mered by assembler. But if I want to use those reads for mapping, is it necessary or reasonable to do that by Perl script?