Denison Riboswitch Detector Help

      A tool for finding riboswitches within a DNA sequence
   The interface | How to make a definition file | How to read the results | Riboswitch references


Denison Riboswitch Detector (DRD) is an online tool that uses a motif-finding algorithm to find riboswitches within a DNA sequence. The specific motif, along with other parameters of the algorithm, can be chosen from a set of predefined motifs or can be changed by creating a new definition file (see How to make a definition file below).

The interface

You will see on the query page that a sample query is set up for you. There is a sequence in the text box and the TPP Riboswitch is selected. If you hit Submit without changing anything, you will see that the algorithm finds two results. Please first try this to get a sense of how the tool works.

Input Sequence: You have the option to either choose a file or copy and paste directly into a text box. (Make sure the correct radio button is selected.) In either case, the sequence must be in FASTA format. The first line may start with a ">" character followed by a description of the sequence. The rest of the text contains the actual sequence, either all on one line or broken up by newlines.

Riboswitch Type: The tool allows a search for one or more riboswitches at a time. If more than one is selected, a separate window (or tab) will open for each one. You must enable "pop-up windows" for this site in your browser to see these additional windows. The parameters of each riboswitch are contained in a definition file, which can seen by clicking on the name of the riboswitch. You may also choose your own definition file; please see How to make a definition file below for more information. To learn more about each riboswitch, visit the Riboswitch references section below.

ORF Parameters: If the "ORF Search?" box is checked, the algorithm will only return riboswitches that are not within an open reading frame. Turing this option off will likely cause the false positive rate of the algorithm to increase. To the right, you can select the minimum length of an ORF, measured in codons.

Submit Query: Hitting the Submit button will create a results page for each riboswitch selected. The program may take up to a minute to run, depending upon the input length and the number of riboswitches selected. However, the results pages will update every 3 seconds to display all the results that have been found so far. Once you see "Your search is complete." in the page header, you will know that the program has finished.

How to make a definition file

A definition file is a simple text file with information about the desired motifs for a search.

The algorithm first breaks the sequence into short segments and searches each segment for a riboswitch. So the definition file begins with two numbers defining the length of these segments and the overlap between segments.

The next line contains the number of motifs in the riboswitch, followed by one line per motif. Each motif line contains the motif itself followed by the minimum number of identities necessary for this motif to be detected by the algorithm.

Following the motifs is one line containing maximum distances between motifs. The first number in this line is the number of bases added to a putative riboswitch before the first motif. The next number is the maximum distance between the first and second motifs, followed by the maximum distance between the second and third motifs, etc. The last number is the number of bases added to a putative riboswitch after the last motif.

The next line contains minimum distances between motifs. The first number is the minimum distance between the first and second motifs, followed by the minimum distance between the second and third motifs, etc.

The next line contains the minimum number of total identities, from all motifs, for a result to be considered a match, followed by the minimum number of identities in a global alignment of Vienna strings, followed by the maximum length of a putative riboswitch.

This is followed by a Vienna string describing the consensus secondary structure of the riboswitch. The last line in the file is this consensus secondary structure with the motifs inserted at the appropriate positions, used to more accurately align the secondary structure of a putative riboswitch with the consensus.

The Vienna alignment score is computed from a global alignment between the consensus Vienna string with motifs inserted and the Vienna string of a putative riboswitch with motifs inserted. The values of the scoring matrix are set as follows, to encourage motifs to align:

match bases10
mismatch bases 5
mismatch base and parenthesis  -3
match parentheses1
mismatch parentheses 0
open a gap -1
extend a gap 0
deletion penalty -10000

How to read the results

When Submit is pressed, work begins and a results window appears for each selected riboswitch describing the progress so far. The pages will refresh every three seconds, leading ultimately to the final results, listed in order of motif identity score, then Vienna identity score. In our tests, matches to previously annotated riboswitches were virtually always displayed on top. Clicking on a result sets up a BLAST query. Below the alignments and scores, optimal and suboptimal secondary structures computed by mFold are displayed with the Rfam consensus image.

Riboswitch references

For reference, we link below to both a primary research article and the Rfam page for each riboswitch family.

Cobalamin article Rfam PreQ1 article Rfam SAM-IV article Rfam
FMN article Rfam Purine article Rfam TPP article Rfam
glmS article Rfam SAM-I article Rfam ykkC-yxkD article Rfam
Glycine article Rfam SAM-II article Rfam yybP-ykoY article Rfam
Lysine article Rfam