This is determined by the method of analysis you are using. I am assuming sequencing, in which case sequencing method used and for NGS methods sequence depth., for typical illumina 100X coverage of bacterial strains sequence error is less than one in 40,000 bases. To look at Bam file alignment to be sure differences are not due to regional low coverage. Whatever your system, you could run replicates of low input from your control strain to benchmark what low level carryover contamination looks like in your system.

i agreement with Frank R Burns

Thank you Frank.