Genome Scale PCR Infidelity Search

1
Genome Scale PCR Infidelity Search
(The Promiscuous Primer Problem)

• Goal An efficient search for the presence of
  potential undesired PCR products that scans
  through 3 billion bases of human DNA.

2
Conditions for Mispriming

• 1. First 3 bases on 3 end of primer must
  perfectly complement 3 bases in sequence.
• 2. Melting temperature (Tm) of the potential
  duplex gt x (annealing temperature of PCR reaction
  y).

• 3. The position of the mismatch(es) between the
  primer and genomic DNA will affect the binding of
  the polymerase, so must account for differential
  binding based on the position of the mismatch(es)
  (Real-Time PCR studies will address this).
• 4. Forward and reverse primers must be within z
  bases of each other.

3
Central Issues of Infidelity Search

• 1. Size of human genome 3 billion base
  pairs.
• ? Run Blast search locally or presort genome as
  prefix tree and write search program.
• 2. Tm Calculations are computationally expensive
  to repeat billions of times.
• ? Precalculate all sequences that will misprime
  rather than calculate Tms on the fly.

4
The Algorithm
?
We know the primer Tm, DH, and DS. Scan all
possible single mutations, and calculate their
Tms going from most to least stable. Repeat
process for two mutations, etc. Once we hit a
threshold value, we know the primer will not bind
with anything else.
5
The Search

• After calculating all the possible sequences
  that will misprime with our primer, we can then
  do a string search.
• The ideal answer
• 1. Run a batch Blast locally with all the
  possible sequences.
• 2. If a forward hit and reverse hit are lt x
  bases apart, we say that is a misprime!
• The hard answer
• 1. Write a search program of the presorted
  genome and hold misprime indices to compare
  forward and reverse primers then check to see if
  they exist lt x bases apart.