Update error models

[HadrienG]

The two error models currently built-in are based on fairly old genomics data. It'd be nice to base them on recent-ish metagenomes, or at least recent genomic datasets sequencing with most popular read-lengths + lib prep

• [x] Identify 1 MiSeq, 1 HiSeq and 1 NovaSeq dataset to use

MiSeq

• [x] Assembly with megahit
• [x] mapping the reads back to the assembly with bowtie2
• [x] build models from the bam file

HiSeq

• [x] Assembly with megahit used public illumina data from human
• [x] mapping the reads back to the assembly with bowtie2
• [x] build models from the bam file

NovaSeq

• [x] Assembly with megahit used public illumina data from human
• [x] mapping the reads back to the assembly with bowtie2
• [x] build models from the bam file

Roadmap reference: Provide at least 1 good HiSeq and 1 good MiSeq model

[HadrienG]

closed with 38c73a1