Chromatin immunoprecipitation and multiplex sequencing (ChIP-Seq) to identify global transcription factor binding sites in the nematode Caenorhabditis elegans

The global identification of transcription factor (TF) binding sites is a critical step in the elucidation of the functional elements of the genome. Several methods have been developed that map TF binding in human cells, yeast, and other model organisms. These methods make use of chromatin immunoprecipitation, or ChIP, and take advantage of the fact that formaldehyde fixation of living cells can be used to cross-link DNA sequences to the TFs that bind them in vivo. In ChIP, the cross-linked TF-DNA complexes are sheared by sonication, size fractionated, and incubated with antibody specific to the TF of interest to generate a library of TF-bound DNA sequences. ChIP-chip was the first technology developed to globally identify TF-bound DNA sequences and involves subsequent hybridization of the ChIP DNA to oligonucleotide microarrays. However, ChIP-chip proved to be costly, labor-intensive, and limited by the fixed number of probes available on the microarray chip. ChIP-Seq combines ChIP with massively parallel high-throughput sequencing (see Explanatory Chapter: Next Generation Sequencing) and has demonstrated vast improvement over ChIP-chip with respect to time and cost, signal-to-noise ratio, and resolution. In particular, multiplex sequencing can be used to achieve a higher throughput in ChIP-Seq analyses involving organisms with genomes of lower complexity than that of human (Lefrançois et al., 2009) and thereby reduce the cost and amount of time needed for each result. The multiplex ChIP-Seq method described in this section has been developed for Caenorhabditis elegans, but is easily adaptable for other organisms.

Variant identification in multi-sample pools by illumina genome analyzer sequencing

Multi-sample pooling and Illumina Genome Analyzer (GA) sequencing allows high throughput sequencing of multiple samples to determine population sequence variation. A preliminary experiment, using the RET proto-oncogene as a model, predicted ≤ 30 samples could be pooled to reliably detect singleton variants without requiring additional confirmation testing. This report used 30 and 50 sample pools to test the hypothesized pooling limit and also to test recent protocol improvements, Illumina GAIIx upgrades, and longer read chemistry. The SequalPrep(TM) method was used to normalize amplicons before pooling. For comparison, a single 'control' sample was run in a different flow cell lane. Data was evaluated by variant read percentages and the subtractive correction method which utilizes the control sample. In total, 59 variants were detected within the pooled samples, which included all 47 known true variants. The 15 known singleton variants due to Sanger sequencing had an average of 1.62 ± 0.26% variant reads for the 30 pool (expected 1.67% for a singleton variant [unique variant within the pool]) and 1.01 ± 0.19% for the 50 pool (expected 1%). The 76 base read lengths had higher error rates than shorter read lengths (33 and 50 base reads), which eliminated the distinction of true singleton variants from background error. This report demonstrated pooling limits from 30 up to 50 samples (depending on error rates and coverage), for reliable singleton variant detection. The presented pooling protocols and analysis methods can be used for variant discovery in other genes, facilitating molecular diagnostic test design and interpretation.