Direct and rapid HiFi reads affiliation to assess putative contamination.
The idea is to do the same thing as with Illumina reads with Kaiju but for HiFi reads. Kaiju does not seem appropriate for the task as it try to find the longest protein hit in the reads while in HiFi reads we have potentially multiple proteins in one reads. Kraken or centrifuge could work as they work with kmers. PacBio has developed a workflow to affiliate and annotate HiFi reads based on diamond and MEGAN LR (pb-metagenomics-tools). This approach requires a lot of computing ressources and is then no suitable in our case where to we just want to assess putative contamination in the reads.