hide
Free keywords:
-
Abstract:
ChIP-Seq couples chromatin immunoprecipitation (ChIP) to ultra high throughput massively parallel sequencing to create in vivo genome-wide maps of protein-DNA interactions. Rapidly eclipsing ChIP-chip, which employs tiling arrays, ChIP-Seq promises higher positional resolution and decreased expense. We generated ChIP-Seq and ChIP-chip datasets for three Arabidopsis thaliana transcription factors: AP2, FD, and SMZ. Coding regions were fused to GFP and ChIP was performed using identical antibodies to precisely isolate differences in DNA binding specificity. All lines recapitulated untagged phenotypes. All experiments yielded high confidence datasets. For example, among the best bound FD targets by ChIP-chip were MADS-gene loci SEPALLATA3 (SEP3), APETALA1 (AP1), and FRUITFULL (FUL). AP2 and SMZ bound several other miR-172-targeted AP2-domain proteins, among which negative feedback regulation was observed, but until now it was unknown whether this was direct. To determine how ChIP-Seq compared with ChIP-chip, we performed systematic comparisons. ChIP-Seq sequencing was performed on an Illumina 1G genome analyzer. Between 2 and 9 million high quality reads were mapped for each sample. With such numbers we obtained >90% coverage to the nonrepetitive genome. Several ChIP-Seq analysis methods were directly compared, including a novel program we designed specifically for Arabidopsis thaliana ChIP-Seq data. ChIP-Seq and ChIP-chip datasets were largely consistent, especially for high confidence(FDR Q<10^-20) binders: for example, FD binding to MADS-gene loci SEP3, AP1, and FUL was detected by both methods with high confidence. Systematic comparison of the relative utility of technical and biological replicates and requirements of genome sequencing depth will be presented in detail.
