Simpler FASTQ_CHECK #102

Our FASTQ_CHECK module now runs fq lint and stops the run with one log message if any of the samples fail. This was implemented on purpose: input and output libraries are scored against each other, so a faulty sample/fastq can lead to (often subtle) biases throughout the entire selection experiment rather than just dropping out. But the module could indeed be further simplified (comment by @maxulysse in #54, with seqinspector as a reference):

• Run fq lint and Falco (instead of FastQC) in one task per library
• Shorter error classification: fewer patterns, relying more on fq's own message
• @andrewdavidsmith is also considering to add (some of) the fq lint checks directly into Falco; if this works we could eventually stick to this alone after updating the nf-core module (https://nf-co.re/modules/falco/)

Planned for our v1.1.0 pipeline performance overhaul.