In bacterial counts, sometimes for the specific bacterial count, such as Enterobactericeace/ Staph. aureus count from milk samples, the count number is null/Zero for some samples. Then what we can do for log transformation of the value. As the log-transformed value of 0 becomes infinitive? And if we remove the samples from the count, perhaps the mean will be changed as the number of observations will be reduced? Would you please correct me or suggest your opinions?

Thanks in advance