To start using PRI-CAT, please go to the 'Data uploading' tab and submit your data.

Look at our TUTORIAL for how to use PRI-CAT, it is very easy.

If you just want to browse our DNA binding maps, please use a DAS-compatible genome browser. We strongly advise the Integrated Genome Browser (IGB):

1)   Launch IGB.

2)   Choose 'configure' in the data access tab

3)   In the data sources tab, add 'PRI-CAT' to Name 'http://www.ab.wur.nl/pricat/quickload' to URL, and 'quickload' to Type.

You will have access to all the DNA binding maps publically available in PRI-CAT.

You can directly download the maps from here. Remember that PRI-CAT is fully compatible with GALAXY. So, you can do genome-wide analysis of your data or any publically available dataset from PRI-CAT using GALAXY (see this for more detailed information).

Please, reference Muino et al. NAR (2011) if you use PRI-CAT.

 

Information about PRI-CAT:

·       Motivation.

·       Objectives.

·       Primary analysis of plant ChIP-seq data.

·       Input data format.

·       How the data is analyzed.

·       Visualizing DNA binding maps.

·       Galaxy.

·       Tutorial.

 

 

Motivation

How genes are transcriptionally regulated? How do transcription factors (TFs) choose their targets? Why do some genes have an enormous change in expression upon the binding of a TF, but other genes with the binding of the same TF are apparently unaffected? Perhaps, the second region is highly methylated? Perhaps, another TF is needed to modify the expression of the gene? If you have dreamed at least once with these questions, for sure you will find something interesting among the tools and datasets included in PRI-CAT.

PRI-CAT is an ambitious project which aims to be a central place for the generation, analysis and visualization of DNA binding maps in plant genomes. PRI-CAT will analyze the raw sequence data obtained by next generation sequencers and provide to the user results consisting in a list of binding sites and potential target genes, as well as, the DNA binding map of the analyzed protein. The result/s can be analyzed in combination with other datasets available in PRI-CAT thanks to its compatibility with GALAXY. All publically available DNA binding maps can be visualized with any DAS-compatible genome browser. At this moment PRI-CAT contains more than 25 binding maps resulting from the reanalysis of publically available ChIP-chip and ChIP-seq experiments, as well as in silico predictions. We are committed to continuing reanalyzing more ChIP-seq data as they are available, but users are encouraged to use their own raw data with PRI-CAT and release their result to the public, or send us their DNA binding maps obtained with other software.

The future success of PRI-CAT strongly depends on you, in particular on how many users will make their DNA binding maps analyzed with PRI-CAT or other tools publically available. Please make publically available your DNA binding maps after publication!!

Back to index

 

Objectives:

PRI-CAT was created with three clear objectives:

1)   Provide an user-friendly environment for the primary analysis of plant ChIP-seq data

2)   Provide a central place for the storage of plant DNA binding maps

3)   Provide the possibility of advanced analysis combining other binding maps/genomic information. At this moment, this is achieved through its compatibility with GALAXY

At this moment we mainly focus in Arabidopsis thaliana, but we are planning to extend our web-tool to other plants. Arabidopsis lyrata and Solanum lycopersicum will be, probably, the next plant genomes included in PRI-CAT. We currently don't have plans to include non-plant organisms.

Back to index

 

Primary analysis of plant ChIP-seq data.

Chromatin Inmunoprecipitation (ChIP) combined with next generation sequencing (ChIP-seq) is a formidable tool to generate precise and accurate genome-wide DNA binding maps of proteins of interest. However, there is a complete lack of web-tools for the analysis of this kind of experiments in the plant field. The computational requirements of the algorithms used for this type of analysis are usually beyond the typical PC computer power that biologist may have. For these reasons, we built our web-tool using our powerful in-house computer servers, so users will not be limited by their computer resources. The simplicity of PRI-CAT is based on our extensive experience analyzing plant ChIP-seq data [ref], that helped us to point out which are the important steps of the analysis that need special attention by the user, and which ones can be completely automated.

We implemented PRI-CAT thinking of the biologist as final user. Users will need to introduce biological information on the system, as for example which is the average DNA fragment size submitted to sequencing, but any other statistical parameters of the analysis will be optimized by PRI-CAT. Users are advised to check the quality of their sequenced libraries with the graphical output generated by PRI-CAT.

The DNA binding maps generated by PRI-CAT report score values based on the ratio between IP and control samples after statistical normalization. Ratios are more suitable for a straightforward comparison of different experiments than Poisson-based scores since they are not so dependent on the statistical power of the test (eg: number of replicates or deep of the sequencing effort). However, users should be careful when comparing experiments using different types of controls (eg: input DNA, IP on mutants, ... ) and when the homogeneity of the samples is very different. Plant ChIP experiments are much more challenging that their animal counterpart since they are usually not based on cell culture. Therefore, plant samples represent complex mixtures of cell types (non homogeneous samples) and this should be taken into account when comparing different ChIP-seq analysis results. PRI-CAT uses the same algorithms (see below) for all the data that it analyzes, which also facilitates the direct comparison of different experiments.

Users are encouraged to make publically available their raw data; especially their control samples which are as expensive to generate as IP samples. Control samples (eg: input DNA) usually don't contain sensible information. However, they can be used by other users to analysis their IP samples without generating an expensive control on their own. PRI-CAT allows users to analysis their own raw sequence data in combination with the data publically available in our server.

Back to index

 

Input data format.

PRI-CAT accepts input data in fasta or fastq format. The files should be compressed before submitting them to the server. Accepted compression formats include .zip, .tar .gz and .tar.gz. There is a limitation of the data uploading size of 3 GB. This usually represents 2-3 ChIP-seq experiments, including controls. In case that the user wants to analysis larger group of files, he is advised to directly contact us for a more efficient way to transmit the data. Sequencing files formats provided by genomic center facilities and private sequencing companies usually can be directly used in PRI-CAT.

Back to index

 

How the data is analyzed.

The sequencing files submitted by the user are mapped to the appropriate genome using SOAP2 (http://soap.genomics.org.cn/) allowing a maximum of two mismatches. Sequences mapping multiple positions, mitochondria or chloroplast regions are discarded. Enrichment is detected using CSAR, an R package specially developed to analyze plant ChIP-seq data. The parameter used were:

For SOAP2: -r 0 -n 20 -l 30 and default parameters

For CSAR: backg=10 norm=-1 test='Ratio' times=1e6 digits=2 considerStrand='Minimum' npermutations=5 and default parameters

We chose SOAP2 because it shows a high accuracy mapping short next generation sequences (see table 1), and because of its speed.

 

Table 1. Results for the mapping process of several algorithms. Sequences of 36 bp length were randomly generated form the Arabidopsis genome (TAIR9) using Maq program allowing the simulation of sequencing errors (`mutations`) or not (`no mutations`) . Table 1 summarizes results as the average over 5 simulations.

 

CSAR was implemented specifically for the analysis of plant ChIP-seq data (see Nature Protocols and Plant Methods). It was developed to be robust against PCR-artifacts. Taking the average size of DNA fragments subjected to sequencing into account, the software calculates single-nucleotide read-enrichment values. After normalization, sample and control are compared using a test based on the ratio or the Poisson distribution. Test statistic thresholds to control the false discovery rate are obtained through random permutations. CSAR shows better performance analyzing plant ChIP data than other compared software (see Fig 2)

 

Figure 2. ChIP-seq method comparison. (A) Enrichment of CArG boxes (known SEP3 binding motif) in peak areas detected by the different methods. Libraries representing a SEP3 ChIP-seq experiment were reanalyzed with the default options for the different methods. For comparison purposes, all scores reported by the different methods were transformed in rank scores, being 0 the highest significant peak by each method. (B) Proportion of peaks detected by the different methods with at least one gene differentially express. AP1 ChIP-seq experiment was reanalyzed with the default options for the different methods.  The list of genes which expression is affected by AP1 was download from here, we used the list denoted “Agilent and_or Operon_BH-0h”.

Back to index

Visualizing DNA binding maps.

The output results of PRI-CAT include a list of binding sites and potential target genes of the protein of interest in an Excel compatible format (.csv). At this moment all the output results will have an unique ID to its proper reference, however we are considering to change its format in the coming months. We are contacting TAIR to find a proper way to create stable IDs for each output that can be referenced in a similar way as GEO IDs. We strongly recommend users to visualize their results in a genome browser, in particular in combination of other genomic maps included in PRI-CAT. DNA conservation, CpG islands, or the binding of other TFs can be extremely useful to interpret the user's results. We strongly encourage the use of the Integrated Genome Browser (IGB); it is fully compatible with the DAS server where we store all ours maps. It also present striking advantages compared with other browsers, for example its speed, low computer power required by the server side, and the possibility to do statistical transformations small analysis transformation of the maps on the fly. The possibility to locate DNA sequence motifs on the fly is also an extremely useful tool in order to interpret DNA binding data.

Back to index

 

Galaxy

GALAXY is a framework for tool and genomic data integration. PRI-CAT is fully compatible with GALAXY, and users can interactively download their desired PRI-CAT datasets directly from GALAXY. For more help regarding galaxy, please see this.

At this moment, we have a GALAXY instance running in our servers to exemplify the power of PRI-CAT when is combined with GALAXY. If you want to work with your own GALAXY instance, please download and install this tool in your GALAXY instance (see readme inside for installation instructions).

Back to index

 

Enjoy PRI-CAT!!!

On behalf of the PRI-CAT team,

Dr. Jose M Muino