Chia-pet

The method was first used in 2009 ^[1] to determine the remote ER-alpha binding sites in human breast cancer. Subsequently, it was used, for example, to construct an interactome (a map of possible interactions) mediated by CTCF in mouse embryonic stem cells ^[4] .

ChIA-PET combines the capabilities of methods based on ChIP and 3C ^[5] , expanding the capabilities of each of them. The traditional ChIP method allows one to determine the interaction of a specific protein with DNA and can be used to search for transcription factor binding sites . Using ChIP-seq , binding sites of the de novo protein of interest in the whole genome can be determined. If a protein binds chromatin regions that are far apart on the chromosome but close in space, ChIP-seq can identify each of them, but will not indicate their interaction. Moreover, not all sequences determined by the ChIP-seq method are uniquely mapped in the genome and not all are functional binding sites ^[6] .

The 3C and ChIA-PET methods are based on the theory of proximal ligation ( Eng. Proximity ligation ), which states that the ends of chromatin sites associated with the protein complex, which are nearby, will be ligated to each other more likely than the ends of sites in solution or associated with another protein complex.

Method 3C allows one to determine the spatial structure of chromatin, but does not allow one to determine the interacting protein ^[5] . A significant problem is the need for accurate knowledge of the sequence of interacting loci . This is necessary for the selection of primers for quantitative or semi - quantitative PCR analysis used to determine them. (It should be noted that there can be several loci - candidates for interactions - determined by the 3C method). The ChIA-PET method allows de novo to determine the spatial structure of chromatin bound to a specific protein. That is, on the one hand, it does not require knowledge of the DNA sequence in the interaction region, and, on the other hand, depends entirely on the specificity of the antibodies used.

Experimental Methods

DNA-protein complexes are non-specifically crosslinked with formaldehyde. The sample is exposed to ultrasound , while the DNA molecules are fragmented into fragments and non-specifically bound weakly bound complexes are destroyed. The result is DNA fragments in strong complexes with proteins. Then, using specific antibodies fixed on magnetic balls, chromatin fragments associated with the protein of interest are precipitated . Often the objects of research are known transcription factors ^[1] . The precipitated complexes are removed from the solution by the magnetic balls using a magnet. The isolated complexes are divided into 2 aliquots and oligonucleotide half linkers with a known sequence are “sewn” onto the ends of DNA molecules. In one aliquot - halflinker A, in the other - halflinker B. Both halflinkers contain a site recognized by the restriction enzyme MmeI, and differ from each other by a “barcode” of two nucleotides : CG for halflinker A, and AT for halflinker B. As a result later on during sequencing it is possible to distinguish linkers from each other according to the “barcode”. At the next stage, two aliquots are combined and proximal ligation occurs, as a result of which half linkers are ligated to each other with the formation of full-length linkers. Linkers with AA (CG / CG) or BB (AT / AT) barcodes are considered probable ligation products within the same complex, while linkers with AB (CG / AT) barcodes are considered chimeric DNA ligation products associated with different protein complexes Preparation for sequencing involves the treatment of complexes with the restriction enzyme MmeI ^[8] , which cleaves DNA at a certain distance from its recognition site in the half linker. As a result, at the end of this stage, designs are obtained containing a pair of “tags” ( English tag ) (20 bp each) on both sides of the full linker (38 bp). The resulting fragments are sequenced from both ends ( English PET ). After that, "tags" are mapped to the genome. ^{[7] [9]}

Computer Processing

Computer processing of sequencing results includes 6 modules ^{[7] [10]}

• Linker Filtering
• Tag Mapping ( PETs )
• Tag Classification
• Identification of protein binding sites
• Determination of chromatin interactions
• Visualization and organization of results

Linker Filtering

All read sequences are divided into sequences having readable “barcodes” and unreadable “barcodes”. If the “barcode” cannot be read, then the sequence is “discarded” from processing. If the "barcode" can be read, then the sequences are aligned by linkers. All obtained sequences are divided into chimeric (formed by ligation of DNA from different complexes and containing A / B linker) and non-chimeric (containing A / A or B / B linker). It should be noted that among sequences containing A / A or B / B, chimeric ones may also occur. Next, the sequences of the linkers themselves are “discarded” and the sequences of “tags” (PETs) are analyzed. ^{[7] [10]}

PET Mapping

The sequences obtained in the previous step are mapped in the reference genome. At the first stage, sequences are aligned that are 100% aligned , which can be mapped at one locus (unique) or at many loci. From the remaining sequences, those that contain 1 substitution in the sequence ( English missmatch ) with the reference genome are isolated , they are also divided into unique and sequences with multiple mapping. All other sequences refer to non-mappable. All sequences except uniquely mapped ones are “discarded” from processing. ^{[7] [10]}

PET Classification

Two groups of PETs are distinguished: “self-ligating” ( English self-ligation PETs ) and “interligated” ( English inter-ligation PETs ). "Self-ligating" ( eng. Self-ligation PETs ) correspond to the ends of a single ChIP DNA fragment, they must be located on the same chromosome at a small distance from each other, in the "head to tail" orientation. "Interligated" ( English inter-ligation PETs ) are divided into intrachromosomal (mapped on the same chromosome at a great distance), interchromosomal (mapped on different chromosomes) and ligated in different orientation tags ( English different orientation ligation PETs ) (mapped on the same chromosome at a small distance, but in the wrong orientation or on different DNA chains). The boundary separating “self-ligated” PETs from “interligated” PETs is naturally determined by the length of the DNA fragments obtained at a given “sounding” intensity. In different experiments, it was 3–4.6 Kb. ^{[7] [10]}

Identification of protein binding sites

Self-ligation tags ( self-ligation PETs ) are used to identify protein binding sites. The procedure is similar to that used in ChIP-seq .

Determination of chromatin interactions

In predicting this type of interaction, “interligation” PETs are used .

Visualization and organization of results

Data from the previous steps are entered into databases for storage, processing and possible visualization.

• In this method, the signal to noise ratio is not too large. This may be due to non-specific protein complexes, insufficient specificity of the antibodies used, PCR before sequencing and sequencing errors. All of these factors can make mistakes.
• The second important point: the need to abandon the analysis of a large amount of data. For example, non-unique PETs located in repeats where functional binding sites of transcription factors may also be present ^[11]
• It is impossible to determine which of the proteins of the complex “sitting” on the regulatory site is responsible for the formation of the DNA loop. This requires additional experiments, for example, 3C.
• Specific antibodies are required, in addition, the protein in the complex may be screened and inaccessible to antibodies, in which case the binding sites of the complex will not be determined.
• A reference genome is required.

• Allows de novo to determine both chromatin interactions and protein binding sites.
• Does not require knowledge of the DNA sequence at the protein binding site.
• The ability to adjust the selectivity of the method by selecting specific antibodies either for a specific protein (for example, a specific transcription factor) or for a universally used protein (for example, general transcription factors).
• Compatible with tag-based NGS sequencing ^[12] .

The following computer programs are used in ChIA-PET experiments.

• ELAND Maps ChIP-rich DNA fragments to the human reference genome. [2]
• Eisen software Determines gene expression levels based on hierarchical clustering. [3]
• RepeatMasker In-silico masks repeating elements. [four]
• Monte Carlo simulation Used to evaluate the level of false results. ^[fourteen]
• PET-Tool Software for processing and working with Paired-End di-Tag sequence data. [5]
• ChIA-PET Tool ChIA-PET data processing software. ChIA-PET Tool web site ChIA-PET Tool paper

• Next Generation Sequencing Methods
• Twin End Sequencing
• Determination of chromosome conformation
• en: ChIP-seq
• Discussion of ChIP on molbiol.ru