I am trying to help a group develop a pair of PCR primers that would permit the accurate detection of one specific bacteria strain. The goal is, using a simple PCR kit, to be able to determine whether the studied strain is in the mix.

I am not used to work with bacteria or metagenomes, so I have a few questions regarding the feasibility and best approach to design these primers.

Some options that we are considering:

Option A:

• design primers in a highly variable region of a gene that is often used for bacteria identification (eg: 16S, GroES, GroEL, rpoD, GCAT, gyrB, cpn60, rpoB)

Option B:

• Get all the bacterial genome data that is available
• Identify kmers that are found only once in our strain and never in all the bacterial data, including other strains of the same species
• Look for pairs of those unique kmers that are separated by 100-800 pb in the target strain's genome to design primers
• Do some sequencing of the targeted strain and similar strains to confirm that the primers are unique to the targeted strain
• Test in silico amplification in all other genomes (bacterial AND eukaryotes)
• Test it in the lab (positive and negative controls...)

Of course, we would test in the lab for false positive/negative results.

My questions are:

• Is using a single PCR likely to be powerful enough to detect the presence of one specific strain with low false positives/negatives?
• What approach would you use to design PCR primers that would be specific to one strain?