How to Use
Note: This site requires JavaScript to function properly. An internet connection with a high upload speed is recommended due to the potential size of uploads. IP addresses of submissions are logged.
PELT
To use this application, simply fill in the form and click submit. The Array Type and Genome must be chosen. Only human genomes are available for tiling arrays.
Results are available for the top 1%, 2% and 5% of probes in BED, GFF, WIG and bedGraph format. A .genes file is also produced containing the gene annotations of passed probes, as well as a PDF and HTML files of the results.
Tiling and Promoter NimbleGen arrays are supported, as well as Promoter Agilent arrays. Agilent data must be supplied as normalized GFF files.
1-8 GFF files can be uploaded, up to a total maximum size of 500MB.
Note: GFF files must be tab-separated, not space-separated.
BELT
To use this application, fill in the form and click submit. A total of 3GB of data may be uploaded whether or not compression is used (data file and optional control file combined). However, if compression is used, the files may decompress to a maximum of 4GB each. At least 10,000 tags must be present in each file. Upon submission, a temporary page will be generated to access the status of your job. If your job encounters an error (such as an incorrectly formatted input file), the page will list the details of the error (example). If your job completes processing, a results page (example) will be generated for you, listing statistical information about the BELT analysis at each specified bin size and threshold percentage number and graphically displaying the distribution and intensity of peaks on the genome. Processed data will be available for download as a .zip archive.
Details of the graphical display are available here.
For those genomes for which annotation is supported (human and mouse), the output files for each specified bin size and threshold percentage are formatted as follows (17 column tabular output). The top fields describe the peak, and the bottom fields describe the gene annotation.
chromosome peak peak peak extended target # of peak score p-value
start end enrichment peak enrichment site control peaks
chr1 5653151 5653250 7 9 5653216 5 2.30735 0.00137428
relation to distance gene end gene strand RefSeq Id gene gene numerical
gene from gene end symbol ID
Gene_Desert 321918 5975118 5845456 - NM_015102 NPHP4 261734
For other genomes, only the first 9 columns will be present. More detailed information about gene annotation is available in the Supplemental. All p-values will be zero without a control file, and the # of control peaks will be meaningless.
The files BELT_Output.txt, dataFile_preProcessInfo.txt and dataFile_BELT_summary.txt contain raw output of BELT, including summary statistics and warnings (such as peaks beyond the length of a chromosome). Similar files BELT_Output_WIG.txt, BELT_Output_BG.txt and BELT_Output_Enrich.txt are also generated if a .WIG, bedGraph or bin enrichment output file is requested. The file annotation_log.txt contains warnings or errors from gene annotation (such as warnings that it cannot process mitochondrial DNA), and error_log.txt contains warnings or errors not directly related to BELT if any.
Descriptions of supported file formats are available here. Please note that .gz archived files are not accepted, they must be .tar.gz even though they should contain only one file each.
Options:
- Format: Specify the format of the input files. Both the input file and the optional control file must be in the same format.
- Compression: Select compression format if any. Archives must contain only one file and no directories (two separate archives are required for control and treatment data), and the uncompressed files are limited to 4GB each (compressed files still 3GB combined).
- Bin sizes: The sizes of the statistical bins used for the analysis. Please specify in a list separated by single spaces. Minimum is 50.
- Specify percentiles: Proportions of peaks to exclude by bin enrichment. For example, a percentile of 0.90 will cause BELT to retain the top 10% of peaks.
- Specify target FDR: BELT will attempt to guess a percentile (from 0.85 to 0.999) at each bin size specified to achieve the desired FDR. The FDR must be from 1% to 50%. Requires several passes of BELT. This is much more reliable when a control sample is provided. Note: only three bin sizes are permitted per run with this option due to computational considerations.
- Masked reads ratio: Ratio of genome that has not been sequenced.
- Specify permutation multiplier: ADVANCED FEATURE. Controls the number of permutations used to calculate the FDR. The number of permutations used is (multiplier) × (number of called peaks).
- Reference Genome: Select the genome to which the ChIP-seq reads were mapped.
- Use control data file: Allows uploading of a control file consisting of reads of DNA unassociated with a protein of interest, for improved FDR (false discovery rate) calculation taking into account biases of sequencing or DNA handling.
- Do not perform robust linear normalization between samples: Visible only if custom control data file is chosen. By default, BELT compares the peak scores of the control data and experimental data and normalizes them.
- Data are from individual with two different sex chromosomes: Select if this is the case.
-
Generate fixed step WIG file for IGB genome browser: If checked, will create a fixed-step .WIG file of the bins enrichment level with the first specified bin size, for use in the IGB genome browser. Requires an extra pass of BELT.
- Format is four-column tabular output by chromosome, start, end, enrichment, with the proper .WIG header lines.
- eg.
chr1 5000 5050 12
- Generate variable step WIG/bedGraph file for UCSC genome browser: If checked, will create a .WIG or bedGraph file (choose the format) of bin enrichment with the first specified bin size. Requires an extra pass of BELT.
- Generate a password to access job: If checked, will return a password upon submission necessary to access your results. The username will always be "BELT" (without quotes). Note: This does not perform encryption.
- Email address: If specified, an email will be sent when your job has finished processing or has encountered an error.
- Help us improve BELT: Grants permission for your uploaded files to be retained, for use only in improving BELT and fixing any bugs that may be found. Files will be held confidential.
Currently supported genomes:
Species | Genome |
H. sapiens | hg18 |
H. sapiens | hg19 |
M. musculus | mm8 |
M. musculus | mm9 |
Experimental: |
C. elegans | ce6 |
D. melanogaster | dm3 |
A. thaliana | tair9 |
Note: Shockwave Flash is required for the BELT Flash-based uploader.
Uploadify is used by the Flash-based uploader, and is available under the MIT license. The graphical display applet in BELT results pages requires
Java 1.6 or later, although results files are still available for download without it.
The applet's security certificate must be accepted in order to save images from it, although the applet will otherwise still function if the certificate is not accepted. This display may take some time to load and refresh, since it must download a portion of the results and generate the graphics on demand. Jobs will be deleted one week from the time of submission.