Over 15 years ago, Scott Hansen at the University of Washington School of Medicine wanted to analyze replication timing for specific loci. He developed an assay that has provided a wealth of information about specific loci and that is still used today. Taking advantage of the ability of next-generation sequencing to extend many analyses to the whole-genome level, Hansen collaborated with John Stamatoyannopoulos's group (and has since joined the lab) to develop the second generation of the replication timing assay.

In this assay, exponentially growing cells are pulse-labeled to 'mark' newly replicated DNA in each cell. The label, 5-bromo-2-deoxyuridine (BrdU), is a base analog that DNA polymerase incorporates in place of thymidine and can be recognized by an antibody to later identify the newly replicated regions. Next, staining with the fluorescent dye DAPI labels the entire cellular DNA, which enables cell fractionation based on the total amount of DNA and thus the cell's stage in the synthesis (S) phase of cell division. Newly replicated BrdU-labeled DNA is then isolated by immunoprecipitation from each fraction.

Previously, researchers analyzed loci of interest one by one, by PCR—a process with limited throughput. Now, Hansen explains, “we basically used the same technique I had been using, except the DNA fractions were [assembled into separate Illumina sequencing libraries and] sequenced.”

After a control experiment showed that the data generated with this 'Repli-Seq' approach agreed with the results of PCR-based assay, the researchers went on to compare early- versus late-replicating DNA in multiple cell lines—from fibroblasts to human embryonic stem cells. They performed a high-resolution analysis, examining six fractions per cell line that represented the G1b, S1, S2, S3, S4 and G2 stages of the S phase. Early replication regions were associated with gene density and chromatin accessibility, but a comparison to microarray RNA-level data revealed that gene expression was only moderately correlated with early replication.

“Given what was known,” says Hansen, “there was a thinking that most of the genome was going to be replicated at about the same time in different cell types; that only a few spots here and there will change.” It turned out, however, that in this analysis of just four cell types, there were replication timing differences in 49% of the human genome. “We expect now that number will go up as we include more cell types,” he adds.

And as next-generation sequencing becomes more accessible, others can use Repli-Seq to analyze their cells of interest—other human cell types or cells from other species; all one needs is a reference genome. For a lower-resolution, more cost-effective analysis to look at just early versus late replication, for example, barcoding could be used, enabling analysis in a single lane of sequencing, suggests Hansen, adding: “We think now we can get the basic profiles of the main S-phase fractions in 1–2 lanes of sequencing.”

An important advantage of Repli-Seq that will permit analysis of numerous cell populations, including rare ones, is the lower number of cells required—about 2,000 per replication time fraction. Such amounts can be obtained from a tissue sample, and eventually Hansen and Stamatoyannopoulos aim to examine entire lineages. These data might one day be among the reference information for many cell lines; for now the group is negotiating to upload their material into the University of California Santa Cruz Genome Browser, with much more data yet to come.