One thing you could do is to place your forward primer end right over the mutated region ( depends on how long mutation is of cause) it should reduce the rate of false positives dramatically. Even better, you should use different approach, detecting mutation directly: Cel1 assay. Check it out. It consists of PCR over mutated region and enzymatic detection of + or - mutation. Hope it helps.

In my opinion Sarah, you need a better strategy.  Your PCR analysis of the wild type gene appears perfect.  To be sure, send the products for sequencing. Design primers that flank the sequence that is knocked in or out.   Sequence these products.  Blast all primers and try to optimize melting temperatures. 

Hi Sarah, in the end your problem is to remove the unspecific from your genotyping PCR. It looks like your primers to detect the mutation are quite bad, but you have realized it and tried to design a new one, still without much success.

If I understand it correctly, you are using a common reverse primer and two different forward primers, one for the wt allele, one for the transgenic (which I guess is a knock-in, and it is mapped). The wt primer pair works well, but the mutant does not.

Now, first of all, since you have two to three bands in the mutant PCR, or bands that disappear and others that appear changing annealing temp or genotype, you should always load a 100 bp marker to keep accurate track of the size of the bands you are obtaining, as you do only in the last gel. Secondly you say your mutant band is ca. 200 bp... it would be good to have an exact size, since in your last get the mutants show a lower band close to 100, while the wts show a higher band close to 200 bp. So, my impression is that with the primers you have, you are not detecting the mutant at all. It happens, primers can be difficult at times.

So, options. Dmitry's option is good but laborious for standard genotyping. I would invest in designing primers that might work. Some guidelines. First, the higher the melting temperature of the primer, the more chances that it will be specific. This does not mean designing long primers, a 18-19 nt long primer with 50-60% GC content will easily work at 58-62 °C. Secondly, the 3' side of the primer is the most important one, because that's where the polymerase binds and starts processing. Best primers have at least 3 GC pairings in the first 5 bases at the 3' and the last 2 are always two Cs or Gs (for example: ...CATGG-3', ...TGACC-3', or even ...GCTCC-3'). The deltaG of the 3' end should be between -12 and -15. To avoid primer dimers between the two primers, both forward and reverse primer should end at the 3' with the same two nucleotides (two Cs or two Gs). Test your primers for secondary structures with any primer analysis software and check that there are no structures with deltaG < -4. I also normally check that the last 7 bases on the 3' side of my primers do not anneal more than 3-4 times on the genomic sequence of my gene of interest (with 7 bases only you will almost always anneal in more places than your specific one, but as long as they are few and distant you don't risk unspecific amplification...). There is anyway no guarantee that primers will work, you will always have to fight a bit  with genotyping PCR!

Anyway, you have your original mouse as positive control, with both the wt and the mutant allele. And you have wt mice as negative controls for your mutant PCR. Always run them both in any screening PCR you run, together with the water ctrl (which you might have noticed in your last gel shows a slight contamination...), this is to check that the reagents you use are clean, if not you should replace them! I also prefer to run the 3-primer PCR instead of two separate PCRs, but for this you have to first have the single ones working and then try to put them together. With a thoughtful primer design it works.

Good luck!

Thank you all so much for your help!!  :)  I will work on designing new primers that will work better.  Thank you!! 

While PCR (with redesigned primers as mentioned above) is a good 'pre-screening' method it may not rule out artifacts. Would cost you a lot of energy if you later find out that either the reviewers would demand you perform Southern or Northern blotting to show an effect on RNA and or proof correct DNA inserting. So better to confirm these now. At least that is what I would do!

Definitely well put Johannes... PCR screening is used routinely to screen mouse colonies since it is quicker and less elaborate, maybe less expensive, but when it is set up it must be verified by a Southern blotting or, when applicable, a Northern blotting!

In my opinion do try to blast your primers by NCBI primer blast or UCSC BLAST program....check if they are binding to more than one place or not.Also repeat this process for newly designed primers.

What exactly is the mutation? If it's a small insertion or deletion, you can detect those by melt-point analysis of your PCR products. Or you can try to design a restriction-digest based assay. OR you can PCR amplify the region with primers that flank the mutated region and Sanger-sequence.

Good luck!