I am attempting to amplify a 15.4kb sequence from genomic DNA. The DNA template is high quality (fragment size >80kb) and I'm using the Generuler HR ladder. Gel is 0.4% agarose. I'm using the enzyme Herculase with 4% DMSO. I designed the primers with primer3 and checked for specificity using UCSC in silico PCR and Primer-BLAST.
Cycling conditions:
Initial denaturation: 92°C, 2 min
Denaturation: 92°C, 20s
Annealing: 63-68°C, 20s (gradient in image)
Extenson: 68°C, 8 min
repeat 10 cycles as above, then 25 cycles adding 20s / cycle to extension (so 8:20, 8:40 etc.), so 35 cycles in total
Final extension: 68°C, 8 min
There are 2 bands that approximately correspond to my expected length, but there's also a 20kb and a 24kb band and shorter band (~10kb). The top annealing temperature of 68°C (5th lane in image) looks best in terms of specificity.
How can I eliminate the unspecific bands? Also, I'm not sure that my product has got the correct length. The template contains many repetitive sequences and GC-rich areas. How can I avoid formation of secondary structures?