ApolloRNA: user guide

Computing and viewing predicted secondary structure of a selected sequence (RNA menu),
Searching and viewing RNA/RNA interactions (RNA menu),
Visualizing quantitative variables all along the loaded sequence (Analysis menu),
Getting information (G and C percent ...) on selected sequence (Analysis menu),
Exporting chained views of a user-defined sub-sequence (File menu).

To get more information about Apollo, see the Apollo User guide.
Examples data files: NC_000868.xml (Pyrococcus abyssi Genbank annotations), NC_000868_mfe150.txt (Pyrococcus abyssi free energy on a 150 nucleotides sliding window)

The RNA menu is automatically enabled when required external softwares are launchable on the computer.
Secondary structure prediction (first item) is activated if RNAfold (Vienna RNA package) and convert (ImageMagick) are launchable.
RNA/RNA interaction search (second item) is actived if RNAfold and RNAeval (Vienna RNA package) are launchable.
Go to Download page of ApolloRNA web site, for a detail download and install protocol of these softwares.

This extension allows the computing and the visualization of the secondary structure of a subsequence, predicted with RNAfold. The user can save it in a image file. convert program of ImageMagick is used to translate the image generated by RNAfold to GIF format image and to flip the image. Note that this functionality is not available if no sequence is loaded in Apollo or if the length of the selected sequence is higher than 1 500 nt.

Select a feature or a sequence, open RNA menu and select "Secondary structure compute...". A window opens (see below). The free energy of the predicted structure is displayed on the right. Note that if more than one sequence are selected, only the structure of the first sequence is computed.

Check the "Corresponding genomic sequence +/- x bases" checkbox and put a x value in the textfield, and the corresponding sequence and structure will be computed and displayed in the same window.
Click the "Save structure as..." button to save the currently displayed image as GIF file.
Click the "Save sequence as..." button to save the currently displayed sequence(s) as a multiple-FASTA file.
Check the "Follow external selection" checkbox and the window will follow selections made in the main Apollo window.
Click the "New structure window..." button to launch a new structure window from the current one.
Click the "Flip image" button to flip image in the vertical direction.

A request sequence is defined and research genomic regions are scanned to find potential RNA targets of the request sequence. Note that this functionality is not available if no sequence is loaded in Apollo or if the "Reverse complement" option is checked.

Select a sequence. (Optional)
Open the RNA menu and select "Search RNA/RNA interaction".
A window for configuring the interaction research opens (see below). If a sequence is selected, some fields of the "Specifications of the request sequence" part are filled.
In the "Specifications of the request sequence" part, specify the request sequence by checking one of the radio buttons. Three ways allow to specify it:
- by selecting a sequence in the main Apollo window before clicking the "Search RNA/RNA interaction" item,
- by entering the low genomic position, the length and the strand of a sequence,
- by entering directly a nucleotidic sequence.
In the "Specifications of search regions" part, select and specify the criteria characterizing the explored regions.
All the selected critera are combinated to determine research regions. See below for examples of criteria and underneath the corresponding research regions painted in gray (start codons are painted in green, stop codons in red):
- Genomic positions 500 to 1500 AND Strand +
- Genomic positions 1 to 2000 AND Defined regions 100-500,1000-1500 AND Strand +
  "Defined regions" syntax: a region is represented by start and end positions, separated by a dash (100-500). Two regions are separated by a comma.
- Genomic positions 500 to 1500 AND Around start codon: -10 nt AND Strand +
- Genomic positions 1 to 2000 AND Strand + AND Intergenic regions
To compute research regions, annotated genes are scanned. Note that the genes of circular genomes which start before the genomic stop position and which stop after the genomic start position are not scanned.
In the "Specification of the search procedure" part, click the tab of the choosen method and enter the method parameters.
The first tab is used to launch the interaction research method based on the fold. A click on the "Info: compute number of windows" button computes and displays the number of explored potential targets. This number is clearly related to the search procedure required time. The second tab (see below) is used to launch the interaction research based on the alignment. A click on the "Save parameters" button saves the current method parameters: they will be reloaded at the next opening of the window.
- Description of the Fold based method. The general idea of this method is to build a temporary sequence which appends the request sequence, a special sequence called linker and a potential target. This temporary sequence is then folds with RNAfold. The interactions between the request sequence and the target are extracted. The free energy of the interaction is computed with RNAeval and the interaction is evaluated throught this value. More precisely, the method explores regions of genome with a sliding window. At each step, the method considers the value of the interaction as explained above between the request sequence and the one corresponding to the sliding window. Only the better potential targets are stored.
  The parameters of this method are:
  - the size of the sliding window (between 5 and 100 nucleotides)
  - the shift of the sliding window at each step (between 1 and 20 nucleotides)
  - a sequence of nucleotides called the linker (this sequence has to be sufficiently long to allow folding, between 3 and 10 nucleotides, and composed of nucleotides that will not be involved in interactions (nucleotide N is prefered but A, U, G, C, T are also allowed)
  - the maximum number of potential targets that are stored (between 5 and 1000).
- Description of the Alignment based method. The general idea of this method is to align the request sequence and a potential target using a local alignment algorithm. The request sequence is aligned with each defined region using the Smith and Waterman alignment algorithm adapted from JAligner [Ahmed Moustafa, JAligner: Open source Java implementation of Smith-Waterman, http://jaligner.sourceforge.net]. The default scoring matrix is defined to look for an interaction between the two sequences instead of an alignment. Thus the A-T, G-C and G-T matchs have positive default values. The method extracts one potential target per region. Targets which have the best alignment score are stored.
  The parameters of this method are:
  - the scoring matrix (values between -50 and 50)
  - the opening gap penality (between -50 and 0)
  - the extension gap penality (between -50 and 0)
  - the maximum number of potential targets that are stored (between 5 and 1000).
  The score of a mismatch not specified in the scoring matrix (for instance A-N) is equal to -4.
In the fourth part, click the "Search" button: the research of targets in the specified regions starts.

When the research is achieved, a new window opens (see below) with a table listing the predicted targets of the request sequence.

The table presents one target per line with the following description of each column:

Position: low genomic position,
St: strand,
Interaction: interaction visualization, (The target is painting in red, the request sequence in blue)
Clicking on the picture opens a new window to better see the interaction,
Prev: preceeding gene and its distance with overlapping gene(s),
Gene Pos: low genomic position of overlapping gene(s),
Gene: overlapping gene(s),
Gene Desc: description of overlapping gene(s) product, and in parentheses the related comments when the product is an hypothetical protein. In a Genbank file, the gene product is contained in its qualifier "/product" and the related comments in the qualifier "/Note".
Next: following gene and its distance with overlapping gene(s),
Strands: respectively, strands of the previous gene, overlapping gene(s) and the next gene, ('>' for strand +, '<' for strand - and '.' when no value)
L: length of the interaction,
Var1: main variable of the search method (the one of the fold based method is the interaction free energy computed by RNAeval, the one of the alignment based method is the score of the alignment),
Var2: secondary variable of the search method (the one of the fold based method is the RNAfold free energy),
N°: order number according to the search method,
Pos N°: order number according to the genomic position.

Click the "Save to text" button to save the interactions in a text file.
Click the "Save to fasta" button to save the target sequences in a multiple-fasta file.
Click the "Save to GFF" button to save the targets in a GFF file.
Click on an interaction picture: a new window opens (see below). Click the "Save image as..." button to save the interaction visualization as a PNG file.

The two previous windows contain a "Configure visualization" button, that allows to modify the characters linking two paired nucleotides in the visualization.

Click this button.
In the new opened window (see below), change characters and click the "OK" button: the interaction visualizations are updated using the new characters.

Apollo users could only display the graph of "GC%" variable on a sliding window. This extension allows the visualization of up to five additional graphs on Apollo's main window. These graphs could be based on internal computed values or from other software outputs.

Open the Analysis menu and select "Show quantitative variable...".
A new window dedicated for quantitative variables visualization opens (see below).

The window is composed of two tables. .

The first one is used to load external data. For each graph to be shown, there are five fields:
Quantitative Variable: name of the loaded file that contains the variable values.
Show: checkbox used to show or hide the graph when it is created.
Color: current graph color. You can change the color of the graph by clicking on it.
Load Data: button that lets you choose a file to be loaded. It also creates a new graph and make it visible.
Delete: button used to remove a graph from the current visualisation.

Note that you can only fill the table incrementally (from the top position to the bottom one). Then all actions (change color, delete or hide/show) are disabled until data is loaded and graph is created.

The second table lets users to visualize computed sequence composition variables graphs. Its behaviour is quite similar from the table above. Each row has eigth fields.
The first three of them are similar than previously, the only difference is that the variables names are fixed. A computed variable can only be set with the help of a reading frame, which determines an area where calculation will be made for a particular position. Fields named Frame and Size are two different ways to set the size of the sliding window. You can make the slider move to a specific size or input this size directly in the next field of the row. Changes are automatically reported to the visualisation panel (Beware that it is sometimes necessary to growth zoom to be able to see the graph, depending on the size of the overall sequence).
The last three fields indicate the mean, min and max values of the corresponding variable.

This is a view of one loaded variable and one computed one:

This extension allows to get information:

G and C percent,
number of non A,T,G,C nucleotides,
intergenic percent,

on selected sequence(s) and the entire loaded sequence.

Select a feature or a sequence, or several ones, open Analysis menu and select "Selected sequence information...". A window opens (see the window below where two tRNAs are selected).

Check the "Corresponding genomic sequence +/- x bases" checkbox and put a x value in the textfield, and the corresponding sequence and structure will be computed and displayed in the same window.
Click the "Save sequence as..." button to save the currently displayed sequence(s) as a multiple-FASTA file.
Click the "New info window..." button to launch a new information window from the current one.
Check the "Follow external selection" checkbox and the window will follow selections made in the main Apollo window.

Apollo users could only print in a file the current view of the main Apollo window. This extension allows the export in files of a set of chained views of a subsequence. The user enters the positions of a subsequence, a zoom factor and calls the export: a set of image files is created. The images are ordered, zoomed and chained views of the subsequence.

Open the File menu and select "Export images". (Note that this option is not possible for the synteny). A window opens (see below).
In the first part of the window, enter parameters of image contain: start and stop positions of the wished genomic subsequence, zoom factor, overlapping between two consecutive images.
In the second part, using the "Browse to change" button, select the output directory and enter the prefix of the name of resulted files.
In the third part, click the "Export" button.

Click the "Compute number of files" button to know the number of files, which will be created according to the current parameters.

Created images are chained zoomed views of the subsequence which starts and stops at the specified positions. Overlapping value determines the number of nucleotids displayed in both consecutive views. Created images are PNG files ; their name is the entered file name, followed by the character "_" and the number of the view.

The genomic sequence of Pyrococcus abyssi GE5 (1 765118 nt) is loaded in Apollo. The export of a subsequence starting to the position 198 000 and stopping to 210 000, with a zoom factor equals to 200.0, with 1 000 overlapping nucleotids in files whose name is prefixed by "NC000868" is loaded. 2 image files are created: NC000868_1.png and NC000868_2.png (see below). Note that the last image, NC000868_2.png, doesn't finish to the stop position, but "contains" the stop position.

view window