ChIP Overview
From Genome@Yale
ChIP, short for Chromatin Immunoprecipitation, is the most frequently used method in the lab. As the term suggests, the method involves immunoprecipitation of protein-DNA complexes from formaldehyde crosslinked cells. Cells are grown under normal conditions and are crosslinked in 1% formaldehyde for 10-15 minute just prior to harvesting. Formaldehyde is a general zero-length crosslinker that reacts with free amines on proteins and DNA. Formaldehyde treatment covalently links the protein to DNA so that these complexes do not disassociate during subsequent biochemical steps.
The nuclei from the crosslinked cells are isolated, and chromatin from the nuclei are fragmented and released into solution by sonication. Sonicated chromatin is cleared by centrifugation and is incubated with the antibody that recognizes protein that one wants to investigate. The antibody-chromatin complexes are purified using beads that specifically recognize the antibody. The purified chromatin is released from the antibody by denaturation. Protein-DNA crosslinks are reversed by incubation at 65C overnight. After reverse crosslinking, proteins are digested by proteinase K. The digested proteins are removed from the purified DNA by phenol extraction. DNA is further purified and subjected to ligation mediated PCR amplification. The amplified DNA is labeled and hybridized to DNA microarrays together with genomic DNA that has not been subjected to purification steps.
There are several microarray platforms available for ChIP-on-chip. These arrays contain either PCR products or oligonucleotides. First generation human ChIP-on-chip arrays contained only promoter sequences. The promoter arrays are used to find direct targets of transcription factors and to reconstruct cis-regulatory network. Second generation human ChIP-on-chip arrays covered small contiguous regions of the genome. Currently the entire human genome can be represented by high density oligonucleotide arrays and are available commercially (entire human genome arrays are available from Agilent, Affymetrix and NimbleGen). On the arrays, the binding sites can be simply identified by their higher signal for the ChIP DNA channel than the reference unenriched DNA channel (red spot in the figure).
Alternative to these microarray-based detection methods, the ChIP enriched DNA can be identified by direct sequencing of the DNA (ChIP-Seq). ChIP-Seq is now the preferred method for detecting protein-DNA interactions throughout the genome for its relative simplicity and higher precision in determining the binding sites.
We use open source programs to analyze the ChIP-on-chip data: Perl, R, and MySQL. We use MySQL to store and retrieve the data. Perl is used to parse, filter, select and clean up the data tables. R is used for statistical analysis and visualization of the data. Standard analysis of ChIP-on-chip data involves intensity based normalization using loess for fluorescent intensities for Cy5 and Cy3 channels for each arrays. If there are replicates, we perform additional between-array normalization, and these arrays are weight averaged. Depending on the array design and platform, we also use error models to derive confidence values for each probe for their enrichment for the ChIP DNA. The normalized, averaged array data are then subjected to statistical filtering. These step are performed using R and Perl. The raw data, normalized signals, and final "hit" lists are then stored as database tables. Additional analysis can be performed to define binding motifs and to determine overall distribution of binding sites along chromosomes and relative to genes. We also analyze these binding sites against other known features within the genome, such as local chromatin structures and DNase I hypersensitivity.
