Software tools for heterozygote identification

Sequence variant identification by visual inspection of the melting curves is surprisingly easy with the appropriate software tools. The process is demonstrated in Figure 9.3 (A-D) for a 179 bp product of the cholesterol esterase transfer protein gene. The original high-resolution data is shown in (A). The wild type sequence is shown in black with different sequence variants in various colors. At low temperature before any DNA melting transition, fluorescence decreases with increasing temperature. This linear decrease in fluorescence with increasing temperature is a property of fluorescence unrelated to DNA melting.

To normalize the data, two linear regions are selected, one before and one after the major transition. These regions define two lines for each curve, an upper 100% fluorescence line and a lower, 0% baseline. The percent fluorescence within the transition (between the two regions) is calculated at each temperature as the distance to the experimental data compared to the distance between the extrapolated upper and lower lines. The normalized result is shown in (B).

The shape difference between genotypes can be made clearer by eliminating the temperature offsets between samples by shifting the temperature axis of each curve so that the curves are partly superimposed. This is usually done at the high temperature homoduplex region (low percent fluorescence) of the curves so that heteroduplexes can be identified by their early drop in fluorescence at lower temperatures, as shown in (C).

Different genotypes are most easily distinguished by plotting the fluorescence difference between normalized and temperature-shifted melting curves. A reference genotype is selected and the difference between all other curves and the reference is plotted against temperature, as shown in (D). Genotypes other than the reference trace unique paths that can be easily identified visually. Difference plots should not be confused with derivative plots often used to visualize probe Tm values for genotyping (Lay and Wittwer, 1997, Bernard et al., 1998). Although there is some visual similarity, the plots and their applications are very different.

Visual inspection is convenient when there are only a few samples to study. However, when there are many curves to compare, intuitive visual clustering becomes less attractive than automatic clustering. Classical hierarchical clustering can identify different genotypes after high-resolution melting analysis (Zhou et al., 2005). Although automatic clustering algorithms can be used, it is wise to always visually look at the data, especially when multiple domains are present. Commercial software available with the HR-1™ and LightScanner™ instruments incorporates these basic analysis components.

Original 100

tu o

Figure 9.3

tu o

Used for normalization

Step 1. Normalized 100

Used for normalization

Step 1. Normalized 100

Software-aided analysis for mutation scanning. Thirty-two different human genomic DNA samples were amplified in the LightCycler® and analyzed on the HR-1™ instrument. The original melting data (Original) is normalized (Step 1) and temperature-shifted (Step 2) to compare the shape of the melting transitions. Differences in shape are most easily seen on difference plots (Step 3) of the normalized and temperature shifted data. Wild-type samples are in black, with sequence variants in different colors.

Step 2. Temperature shifted 100-

Step 2. Temperature shifted 100-

Step 3. Difference 32

24e c

Step 3. Difference 32

24e c

86 88

Temperature (°C)

86 88

Temperature (°C)

Was this article helpful?

0 0

Post a comment