DNA Melting - Help page

stitchprofiles.uio.no

Introduction

This help page contains usage instructions and documentation. Helpful explanations of melting concepts can also be found in the glossary. For general info about this website, conditions of use and litterature references, see the about page.

DNA sequence input

You must specify a DNA sequence in one of three ways (remember to click the radio button):

  1. Upload a sequence file. The file must be a text file containing the sequence in the following format: Any of the letters a, c, g, t, A, C, G and T is taken to be one of the four DNA bases, while everything else in the file (spaces, numbers, other letters/symbols) is ignored. The bases encountered are put together as one sequence. The FASTA format is currently not supported, although some FASTA files might actually be interpreted correctly this way.
  2. The sequence's GI number in the NCBI GenBank. A GI number is a series of digits that are assigned consecutively by NCBI to each sequence it processes, but it is NOT the same as the Accession number (or the version)! For a discussion, see NCBI's notes. Examples of GI numbers are: 2423444 (a 360 bp cDNA from C. elegans) and 50878110 (a 14664 bp small bacterial genome).
  3. Insert DNA sequence (copy/paste). Letters a, c, g, t, A, C, G and T are interpreted as the four DNA bases of a single sequence, while anything else (spaces, line breaks, numbers etc.) is ignored (as in (1) above). An example of allowed input is:

            1 tttaatatat ccattcacat tcctataagc taacaagtta ttatccgatt accacaccat
           61 attatgctaa atctaagcaa taagtgtaac gagcttactc gtacaacttc ataatagctg
          121 atatgttatt taactttttt ctagcttctt tttactatgg aaaaaattaa gcggtaaggt
          181 aagtcaaaaa gtcaatgact caaaaaggca ttaagtcatt gacttaatgc ctttttgggg
          241 caacttgtta tttttgccga attatatgac tgaaagacaa cacaaaactg ttaatttact
          301 ttttaatgag aagtttatta aatatttctt tatgtatact ttcattgttg atgacccttg
Extract DNA sequence fragment (optional):
This option allows you to cut a segment out of a very long sequence, by specifying a start and stop position of an interval [x,y]. This may save computation time. The sequence specified above (1-3) is numbered from 1 to N, where N is the sequence length, and therefore x must be greater than or equal to 1 and y must be less than or equal to N. The fragment [x,y] is then "cut out" to form a shorter sequence of length M=y-x+1, which is then further analysed. Note that this may produce artificial end effects, since the ends of a DNA molecule tends to be more thermally unstable than the interior. Note also that in the results, the bases of the fragment are renumbered from 1 to M.

Maximal sequence lengths:
If the computations take more than 120 seconds, you may get an Internal Server Error message. To avoid this situation, you should try to limit the sequence length. Some guidelines are listed below:
Stitch Profile (for a chosen temperature)
max 140 kbp
Melting Curves
max 480 kbp
Probability Profile
max 1800 kbp
Probability Profile (using "Exact power law" option)
max 11200 bp
Temperature Profiles
max 20000 bp

Stitch profile parameters input

A stitch profile describes the conformations of a DNA at a given temperature. Therefore, you must select one of two possibilities (remember to click the radio button):

  1. Choose a temperature: The temperature should be in °C units. The correct format is e.g. 85.3 while 85,3 is wrong. Remember that usually a DNA melts somewhere in the range 50°C - 100°C (circa).
  2. Choose a helicity: The helicity should be specified as a value between 0 and 1 (i.e. not in percent), for example: 0.75. The server then finds the temperature at which the helicity equals that value (low helicities correspond to high temperatures and vice versa, cf. melting curves), before it begins the stitch profile calculation. If you do not know at which temperatures your DNA melts, then you can always use the default 0.5 helicity, which produces a stitch profile at the melting temperature Tm of your sequence.

Futhermore, two parameters are needed that control the search by the probability peak finding algorithm that produces a stitch profile. You can change the two parameters, the maximum depth and the probability cutoff, or you can simply accept the default values: Dmax=3 and pc=0.01. The peak finding algorithm and the roles of the two parameters are further described in the introductionary article Tøstesen (2005).

Here is a tip about the probability cutoff: If you want to do more versions of a stitch profile, the only difference being their probability cutoff values pc, then you can first submit the lowest pc value to obtain one of the plots. From the results page with that plot, you can then submit another pc-value (which should be greater than the original value). This redraws the plot using the new cutoff, which is faster than going back to the input page to submit a new calculation, and the result is identical. The cutoff simply determines what stitches to include and what to exclude from the plot. In this way, you can also adjust the visual appearence of the stitch profile to make it most informative.

How to specify a temperature range for a melting curve

The helicity is a function of the temperature. In principle, any interval of this function can be plotted, but the interesting part is the melting range of temperatures where the function drops from 1 (everything is helix) to 0 (everything is melted). Usually, the melting range is somewhere within 40°C - 110°C, depending on the sequence and the salt concentration. The server can locate the melting range automatically, if you indicate an interval of helicities (a range on the y-axis instead of the x-axis). You have two choices (remember to click the radio button):

  1. Indicate a temperature interval (the range on the x-axis of the plot). Preferably somewhere within 40°C - 110°C. Temperatures are in degrees celcius.
  2. Indicate a helicity interval (the range on the y-axis of the plot). Use values between 0 and 1, but avoid using values equal to 0 and/or 1, since these may be physically impossible for the molecule to fully reach. The two specified helicities will be translated into temperatures by the server.
When plotting the function, the given temperature range is simply walked through with a certain step size, which you can specify:
Relative steps:
The density of steps can be high, medium or low (selected by the user). The density is then calculated automatically with regards to calculation time (to avoid melting curve calculations that take too long time).
Manually: (advanced)
Here you can directly specify the step size in degrees celcius. There may be a minimum allowed step size.

At present, a calculation of differentiated melting curves, i.e. -dθ/dT as a function of T (a curve with peaks), is not available.

Limitations: Max length of DNA sequence is 48kbp. This calculation will complete in ca. 3min. User should use much smaller sequence to get a good plot: ex. a sequence of length 50kbp can make (480/50) = 9 positions in a plot before going over maximum of 480kbp.

How to specify one or more probability profiles

A probability profile depends on the temperature. A plot of a single probability profile describes the structure at a specific temperature only. In order to get structural information at different temperatures in the melting range, you may plot several probability profiles. The curves will be plotted in the same diagram. You have two choices (remember to click the radio button):

  1. at specified temperatures (°C): Type a list of temperatures in degrees celcius. The list contains at least one value. More values are separated with whitespace. For example, you could type 80 82 84 or just 80.
  2. at specified helicities: Type a list of helicities (values between 0 and 1). The list contains at least one value. More values are separated with whitespace. For each helicity, the corresponding temperature is then automatically found by the server. For example, you could type 0.25 0.5 0.75 or just 0.5.

How to choose a probability level for a temperature profile

The probability level of a temperature profile is a value p between 0 and 1. The plot shows the temperatures at which the different regions of the sequence are basepaired with probability p. If, for example, p=0.5 then the regions are melted and basepaired with equal probabilities at the temperatures shown in the temperature profile. (This special case of a temperature profile is also called a Tm profile or stability map.)  Regions that are more thermally stable have higher temperatures in the profile, and regions that are more thermally unstable have lower temperatures in the profile. A lower p-value gives higher temperatures and a higher p-value gives lower temperatures. The server calculates a temperature profile by first calculating a set of probability profiles at temperatures between 40°C - 110°C, and then interpolating between them. You can use the default 0.5 value, or you can type another value between 0 and 1.

Advanced options

Here you can change some thermodynamic and algorithmic settings, but the default settings are recommended.

Loop Entropy Factor:
For the calculation of the classical melting curves, probability profiles and temperature profiles, you can choose between two versions of the DNA melting algorithm described by Tøstesen et al. (2003) using different methods for handling the loop entropy factor:

  1. Multiexponential approximation (fast): The computation time grows as O(NlogN), which is OK for long sequences. The approximation substitutes the loop entropy factor with a sum of exponential functions as described by Fixman & Freire (1977). The necessary number of exponential functions grows logarithmically with sequence length.
  2. Exact power law (slow): The computation time grows as O(N2) for long sequences, which is too slow for long sequences, but OK for short sequences.
Empiric thermodynamic parameter set:
Thermodynamic parameters describe nearest neighbor stabilities, salt dependence, loop entropy, and so on. The parameters are applied using the scheme described in Tøstesen et al. (2003). The following published parameter sets are implemented on this server:
Salt concentration:
You can specify a salt concentration (i.e. sodium [Na+]) in molar units M. Note that this option is only supported by the parameter sets: (Blake and Delcourt 1998), (Blossey and Carlon 2003) and (SantaLucia 1998). The higher the salt concentration, the more thermally stable is DNA and the higher are the temperatures at which it melts (but the relationship is not linear). The salt dependence is empirical and based on measurements in the range 0.01M - 1.0M. The allowed input range is 0.001M - 8.0M. Gotoh and Tagashira (1981) uses a fixed [Na+]=0.0195 M. Fixman and Freire (1977) did not specify the salt concentration in their paper.

Valid XHTML 1.0!