Tips on oligonucleotide design

Jump directly to oligo length or to specificity.

How to choose a Tm range regarding the oligonucleotide length and the GC content ?

        In order to properly choose a Tm range during oligonucleotide design, we have investigated the relationship between the Tm and both the GC content and the length of the oligo. For each oligonucleotide length comprised between 15 and 70 mer (increasing step of 5 mer), we have generated random oligonucleotide sequences with a GC content in the 30 to 70 % range. As example, for a 15 mer, we have generated 1000 sequences containing 5 G or C (33% GC), 1000 sequences containing 6 G or C, etc. up to 10 G or C (66% GC). In a same way, we have generated 29,000 70 mer sequences with a number of G or C comprised between 21 and 49. The Tm of each sequence were computed using the Nearest-Neighbor model and a DNA and sodium concentration of respectively 1 uM and 1M.

        In the figure below, the Tm of each 15 mer oligonucleotide was plotted. As one can see, the extreme Tm are not representative of the set. Few sequences show a significantly higher or lower Tm than the majority of all other sequences. We have observed the same behavior for every oligo length and GC content. Since these low and high Tm are poorly representative of the set, we have chosen to filter out the 2.5% extreme Tm. These thresholds are represented by the black vertical line on the figure.


        We have applied this filter to every set of Tm. In the graph below, we represent in blue the 1000 Tm for each 15 mer set. The yellow dots represent the extremities of the distribution of the 95% selected Tm. We consider that these values are representing the best the Tm range that you can expect for an oligonucleotide as a function of its lentgh and GC content. The minimal and maximal Tm appear to be linearly corelated to the GC content. This is true whatever the oligo length.


        For each oligo length, we have fitted the yellows dots with a straight line, and this for lower and higher Tm. The equations allow us to compute a theorethical Tm for each GC content comprised between 20 and 80 percent, using an increasing step of 5%. We have plotted these theoretical Tm as a function of the GC content and the oligo length in the 3D  figure below. By this way, we obtain two layers that define the upper and lower Tm that one can expect for a given oligonucleotide. The distance between each layer is only a function of the oligo length and does not move with the GC content.


        If you want to see the space between each layer and if your browser supports Java applets, click at the bottom of the window below and drag the mouse vertically to the upper part. Green balls belong to the lower layer and blue balls represent the upper layer. Balls represent the Tm for a given length (multiple of 5 nt) and a given GC content (multiple of 10%)

alt="Your browser understands the <APPLET> tag but isn't running the applet, for some reason." Your browser is completely ignoring the <APPLET> tag! 

        Now, we can use these data to predict a Tm range as function of the GC content and the oligo length. If your browser dsplays Java applets, you can use the tool below to predict a Tm range. For a given GC content and length range, we report the narowest and the largest Tm range that one can expect. The narrowest Tm range is defined by the upper expected Tm for the lower content and length values and the lower expected Tm for the larger content and length values. In the same way, the largest Tm range is extended to the lower and higher expected Tm.


Optimal oligonucleotide length

oligo length (specificity, active sequence / spacer) 

What Blast database to use for specificity computation ?
Blast database for specificity