It is noted that in microbial world many behaviors of microbes are associated with their population density. For example, some antibiotic productions, biofilm formation or enzyme productions are regulated by quorum sensing system.
1) Screen the genome for gene products that are similar to quorum sensing signals
2) Knock out the candidate genes and assay to see if the "behavior" still exists
3) If you are lucky and get this far, you can then take filter sterilized media from a dense wild-type culture (i.e. contains high levels of quorum sensing molecules) and mix it with your mutant cells to see if you can then stimulate the "behavior".
If you don't have any candidate genes, you could generate a library of random mutants with a transposon and try to identify a "behavior" mutant. You could then map the transposon insertion site and to find out which gene is interrupted (google "inverse PCR"). That's a lot of work though, make sure it's worth trying.