Mutation scanning - to quickly eliminate wild type (normal) sequences
Mutation scanning by high-resolution melting can eliminate over 95% of DNA sequencing in resequencing projects. The vast majority of data generated by these projects are of normal (wild type) sequences, and high-resolution melting has the sensitivity to quickly identify those normal samples. The first step in DNA sequencing is to amplify the target by PCR. Therefore, to insert the step of high-resolution melting prior to sequencing is trivial. Mutation scanning by melting is non-destructive, so any sample suspected of carrying a mutation can be further processed for sequencing if simpler methods for identification (such as
DNA matching, and
genotyping by amplicon melting
) fail.
Closed Tube System: High-resolution melting operates by detection of heteroduplexes (mismatched duplexes) that form after amplification of heterozygous DNA. Some of the older mutation scanning methods also operate on this principle, but high-resolution melting is the only closed-tube method and does not require processing, reagent addition, or separation after PCR. All other methods require some manipulation of the product after PCR, followed by physical separation such as electrophoresis, chromatography, or mass spectrometry (see diagram above). Not only are these older methods labor intensive, but the need to open sample contianers after PCR is a severe disadvantage, increasing the risk of contamination in future reactions because PCR products are exposed to the environment.
Fig 2. Above are melting curve plots useful in mutation scanning. WT: wild type homozygote; HET: heterozygous DNA (A>G single base change).
High-resolution melting uses a heating rate of at least 0.1ºC/sec, and is usually complete in 1 to 5 min. Mutation scanning by this method depends on the melting of heteroduplexes that distort the shape of the melting curve. In order to accurately compare curve shapes, the high-termperature end of the curves are superimposed by shifting the curves along the temperature axis until they are overlaid (see Fig 2A). Because the difference between curves is small, we magnify it by plotting the
subtractive difference between samples (Fig 2B). Each curve is usually subtracted point-by-point from the homozygous reference (or an average of all wild type curves analyzed). Although difference curves look similar to derivative melting curves (Fig 2C), the twoshould not be confused.
Sensitivity and specificity of high-resolution melting analysis is better than many of the conventional mutation detection methods including dHPLC. It is more sensitive than DNA sequencing when variants are present at low allelic fractions (high-resolution melting can detect down to 2% of variant DNA in the background of normal DNA, whereas DNA sequencing can only detect down to about 20% variant). Single-base changes, insertions and deletions can all be detected, as long as the PCR primers flank the variation.
The sensitivity and specificity of scanning for heterozygous single-base changes are 100% for PCR products shorter than 400 bp. In PCR products between 400 and 1,000 bp, sensitivity is 96.1% with a specificity of 99.4%. The position of the variant within the PCR product does not affect scanning accuracy.
Although designed to detect heterozygotes, high-resolution scanning often detects homozygyous changes as well. About 96% of human single-base changes are homozygotes that differ in melting temperature (Tm) and are detectable. What is more surprising is that most homozygotes can be differentiated by curve shape changes alone after temperature shifting has been performed. Even though many homozygotes can be detected, it is still wise to mix an unknown sample with a known wild type sample for detection of hemizygous variants (X-linked or Y chromosome) or if homozygous variants are highly likely.
