add workflow for evaluating potential refDB-contaminations

[jvollme]

• [x] adjust nucleotide database creation to use "parse-seqids"
• [x] extract best hit and best contradiciton from nucleotide referenceDB (assuming it is not eukaryotic)
• [x] blast both aganst nucleotide reference DB
• [x] for each, examine the hits with at least 50% coverage (of the shorter of subject or query, respectively) and 90% identity
• [x] for blastx: require 90% coverage of the subect and 90% identity
• [x] count phylum and domain distribution among filtered hits
• [x] if more contradicting domains/phyla than expected domain/phylum --> assume contig is a contamination