More sequences input, much less it will cluster with refseq207

[ray]

Dear authors：

I test vcontact2 using part of my sequences or the whole sequences. (Though I would expect some small difference of the results) the results differed greatly. More sequences input, much less my sequences would cluster with refseq207.

Is there something wrong with my code：

time vcontact2 --raw-proteins cat_virome.faa
--rel-mode 'Diamond'
--proteins-fp cat_virome_map.csv
--db 'ProkaryoticViralRefSeq207-Merged'
--pcs-mode MCL --vcs-mode ClusterONE
-t 60
--c1-bin /data/db/MAVERICLab-vcontact2-34ae9c466982/bin/cluster_one-1.0.jar
--output-dir est90_only.vContact2-refseq207

Thank you！

[ray]

For example, if input 1000 sequences, 30 sequences could cluster with at least one genome in refseq in a subsectet of 700 sequcences . if input 10000 sequences, only 15 sequences could cluster with at least one genome in refseq in the same subsectet of 700 sequcences. if input 250000 sequences, 2 sequences only. What is more, the 2 sequences are not all in 15, the 15 are not all in 30!