mmseqs easy-search faster?

[LauraVP1994]

Hello, I'm using the following command to find the closest references for multiple reads. As both the database and the fasta file are quite big, it takes quite some time. I tried using max-seqs=7, however, the output is really not good giving results that have only 91% while there are more sequences at 96%? Does someone have an idea how to optimize this command while being sensitive (or is this not possible)?

Thanks!

mmseqs easy-search {FASTA_GI} {REF_DB_GI} \
                --search-type 3 --start-sens 2 -s 7 --sens-steps 3 --alignment-mode 3 \
                --sort-results 1 --threads {cpus_per_task} \
                {TSV_GI_SEARCH}