Alladi Ramakrishnan Hall

How much can we quantitatively infer about transcription factor - DNA binding from sequencing-based experiments?

Vishaka Datta

NCBS Bangalore

The regulation of transcription by the binding of transcription factors (TFs) to the genome was one of the first and most well-studied mechanisms of gene regulation in both prokaryotic and eukaryotic cells. The advent of low-cost sequencing provides a way to simultaneously study the binding of a particular TF near thousands of genes inside cells in a single experiment. These experiments are based on a protocol known as ChIP-seq (Chromatin Immuno-Precipitation and sequencing) that was developed nearly a decade ago.

Each ChIP-seq experiment provides two pieces of information about a TF -- the
sequences of the genomic regions bound by it and a noisy yet quantitative measurement of it's binding affinity at each of these regions. Theoretical studies based on this first aspect of ChIP-seq data have already revealed a wealth of information about TF-DNA binding. However, few studies analyze the noisy read-out of binding affinity provided by ChIP-seq, which has the potential to reveal interesting phenomena that occur during TF-DNA binding.

In the first part of my talk, I will briefly describe an algorithm that we developed to detect one such phenomenon, which are interactions that occur between two proteins that binding adjacent locations on DNA. These interactions are known to shape differences in gene regulation between closely related species and play a major role in diverse biological processes such as the interferon response of immune cells and embryonic development.

In the second part of my talk, I will describe very preliminary work on calculating limits on how much we can infer about TF-DNA binding from sequencing-based protocols such as ChIP-seq. The variation in results between two ChIP-seq experiments carried out, say, in different labs, can come from two sources -- either from biological variation in TF-DNA binding between different cells, which is actually of interest to biologists, or artefacts introduced during certain steps of the ChIP-seq protocol. One of the goals of this work is to compute the relative contributions from both these sources of variation from ChIP-seq data, which would guide the development of more accurate measurements of TF-DNA binding affinities.