TilingScan Tutorial

About TilingScan.
TilingScan is a WebApp for the analysis of differentially expressed DNA regions in tiling microarray data in a strand-specific manner.
TilingScan can load data from tiling array results as genome graphs, which display probe intensity alongside the genomic sequence, aligning it onto the reference genome provided by the user.
It allows customized visualization of either specific chromosomes or specific genes of interest.

By uploading data sets that reflect the ratio between two experimental conditions, peaks and valleys of intensity display differentially expressed regions. Along with their visualization, TilingScan provides the user with a tool to accurately locate and identify these regions, recording useful information such as their chromosomal coordinates, length, and mean intensity value on each DNA strand.
All images, tables and results provided by the application can be downloaded for further analysis.

You only need a web browser to execute TilingScan. No other software is required.
Compatible browsers:
Introducing the app: Dashboard.
When you enter the App, you will see the following image on the screen:

  1. Project search: Here you can search for all the already existing projects that you have previously created. By introducing the Project ID, project name or the name of the author all uploaded data sets will be displayed.
  2. Create new project: Opens the page where you can upload your data to create a new project. You can give it a name and save it. TilingScan will also provide your new project with an automatically generated identifier (Project ID)..
  3. Try the app: If you do not have any data sets yet, you can try all the app features with demo data.
Create new project.
By clicking on Create new project, you can upload the data files needed to use TilingScan.

You must provide the following information:
TIP: You can upload every file compressed in a .zip file.

After selecting you data files, click on Create! to upload them to the server. The size of this type of files is usually large, so this process may take a few minutes.
Once the upload is complete, an automatically generated identifier will be assigned to your project (Project ID). You may want to save this ID for future access to it via the project home. Alternatively, you can search for it by providing the application with either the project or the user name.

IMPORTANT NOTE: You can create an unlimited number of projects, but every project will only be available over 60 days. Sixty days after the date of creation the project will be automatically deleted.

Data formats.
Annotation file
GFF stands for 'general feature format' or 'gene finding format'; it is a tab-delimited file with 9 columns. There are several types of GFF files that use incompatible syntax. The original GFF format is GFF1. A variant called GTF is also used. GFF3 has been proposed to extend on GFF and to constrain the specification more tightly to avoid mutually-incompatible versions of GFF.

The annotation file must be an .gff (Generic Feature Format Version 3 , GFF3) file.
Please see http://www.sequenceontology.org/resources/gff3.html for a detailed description of the Generic Feature Format (GFF).

Experimental files
To upload you data files, create a tabulated text file (.txt) containing the following information:

NOTE: In between sequences, you have to leave a blank space.

Download example files
You can download an example data set here.

Visualize Chromosome
By clicking on Visualize chromosome and selecting your chromosome of interest, you can visualize the expression alongside the entire chromosome. To do so, you must set the desired Gauss filter repetition times (default is 3 times). The more times the filter is applied, the smoother the signal profile will look.

So as to facilitate the visualization, the whole chromosome length is split in sections. You can scroll left/right to move along the section, as well as to go back and forth to the different sections by selecting them in the pull down menu (bottom, left). You can download all generated images, either individually or all of them at once.

Visualize Gene
By clicking on Visualize gene you can visualize the expression of a specific gene. To do so, you have to select:

TilingScan will open a detailed image of the selected gene:

Window Search
Window search is a tool for the detection and identification of differentially expressed regions.
By selecting your chromosome of interest, you will locate and identify them. An array of images containing all detected regions will be displayed.

For all the these regions, TilingScan will automatically register the following features: To start your search, you have to select the following features:

Selection and visualization of manually delimited regions
When you are visualizing either an entire chromosome or a specific gene, you can delimit a region of interest by clicking on the graph. When you delimit a region, it will automatically get highlighted and a magnifying glass icon will appear in the center, along with the length of the region in bp, and the start and end points (bar at the bottom).

By clicking on the magnifying glass icon, you will open the image of the delimited region in a new window, that can be adjusted to your convenience.

Annotate! Is a tool that allows the annotation of manually selected regions of interest. When you delimit a region manually, you can click on Annotate! (purple button at the bottom) to register information about it. TilingScan will automatically register the following features: All this information will be recorded in a table that will automatically open in a new window.

You can edit, delete, and save the information to your convenience. This information is saved automatically in your computer, so newly annotated regions will be added to your list as you select them.

NOTE: This information is not uploaded to the server, so it will only be available in the one computer the analysis has been started at. If you wish to continue with the analysis using a different computer, you can download the list for further use.

Description of the search algorithm.
For the detection of differential expression, the application has been implemented with a search algorithm, that we define as the "sliding window search algorithm".

Firstly, two main parameters have to selected: Once these two parameters are defined, two adyacent windows of the desired V size (A and B) are defined at the beginning of the data set. For each of them, the average intensity value of the probes contained within them is calculated. These two windows will slide along the data set, until the difference between the average value of A and the average value of B surpasses the fold-change threshold U. When this happens, the start point of a region will be defined from the first point of B.

To determine the end point of the region of change, a new window (C), of fixed size = V will be created adyancent to the end of B. The comparison between B and C will now be repeated in the same way it was done for A and B, this time extending window B until the difference between B and C is equal to the selected threshold (U). This will determine the length of the detected region, that will be determined by the start and end point of the extended window B.

To prevent false positives from appearing, we establish two criteria that all detected regions muts meet: All the defined above will serve for the detection of up-regulated regions. In order to detect those regions that are down-regulated, the inverse of the signal is calculated, and the same criteria algorithm with the same permutation criteria is applied.

Analyze NGS data.
TilingScan can be used to analyze NGS data from BAM files. For this, we have developed a program called Cover2Tiling, that converts BAM files into TilingScan compatible files locally, and can be freely downloaded at https://github.com/TilingScan/cover2tiling.

How to use it
To convert BAM to TilingScan input format, first open an terminal and navegate to your data directory. Next, use the SAMTOOLS mpileup command for generate the coverage of your BAM:

Remenber change your_bam_file.bam with the name of your BAM file. This will generate you a large file called cover.txt with te coverage of your data. Finally, execute cover2tiling with the next options:

Where: Now you can use the output file generated into TilingScan.

For further information, please read the complete manual of Cover2Tiling, available here.


Home Tutorial Support Dashboard Annotate! GitHub

© 2014 Vicente Arnau, Jose Enrique Perez, Ana Miguel & Jose M. Juanes

Universidad de Valencia - Spain