That’s a great step forward Superman and team ! Sequence matching is the most fundamental operation in bioinformatics space, so enabling a search against OMXWare sequences is a very welcoming step forward.
Search operations can be divided across hierarchies as you mentioned in your question above. Given that OMXWare contain only bacterial sequences, it would be useful to search the entire space (say the level 0 search). A reduced form of search on taxonomic structure (level 1) can include/exclude certain genomes and search against the rest. It will be very useful to keep the search “interface” or queries, similar to what is considered standard in bioinformatics space, so more biologists can naturally use it without requiring additional learning curves. Some examples as how the community uses sequence searches are here - Nucleotide BLAST: Search nucleotide databases using a nucleotide query and jackhmmer search | HMMER .
I would discourage gene level searches, simply because OMXWare already has APIs to extract gene level information. So users can download the sequence they want to use as reference and do a sequence matching on their own platform. Further, OMXware provides various metadata etc along with sequences, so a user should be well equipped to do things of interest and not overburden the OMXWare compute infrastructure.
In terms of usage, the searches will enable users to ask questions related to sequence alignment and sequence homologs with respect to their input sequences.