Annotation of proteins using UniRef
======

Step 1: Searched Prodigal-inferred protein sequences against both UniRef90
and UniRef50 reference sequences from UniProt release 2019_07 using DIAMOND
v0.9.25. The command was:

```
diamond blastp --index-chunks 1 --id 90 --subject-cover 80 --query-cover 80 \
  --max-target-seqs 1 --threads 16 --db $db --query $input --out $output
```

Step 2: Merged search results. For each protein, took UniRef90 as priority,
then take UniRef50 only when it was missing from UniRef90.