Question 1

Which quality metrics should be checked before analysing a library?

Accepted Answer

Fragment-size distribution should show clear nucleosomal periodicity, transcription start site enrichment should be strong, and mitochondrial read fraction and PCR duplication should be low. Library complexity estimates indicate whether deeper sequencing would add information. Libraries failing these checks should be repeated rather than corrected computationally, since the lost signal cannot be recovered.

Question 2

Why does the transposase insertion offset need correcting?

Accepted Answer

The enzyme inserts adapters as a dimer, so the actual cut site sits a few bases inside the read start on each strand. Shifting reads on the plus and minus strands accordingly aligns the inferred insertion point with the true accessible position. Skipping this step blurs footprints and shifts motif centring in downstream analyses.

Question 3

How should differential accessibility between conditions be tested?

Accepted Answer

A consensus peak set is built across samples, counts are assigned per region, and a count-based model designed for overdispersed data is applied with replicates as the unit of inference. Normalisation choice matters when global accessibility shifts between conditions. Replicate number should be decided during design, since technical replicates cannot substitute for biological ones.

Question 4

How reliable is footprinting for inferring factor occupancy?

Accepted Answer

Footprint depth depends on sequencing depth, factor residence time and the intrinsic sequence preference of the transposase. Weak or transient binders often leave no detectable footprint, and uncorrected sequence bias produces artefactual ones. Bias-corrected methods and orthogonal occupancy evidence should support any claim about a specific factor at a specific site.

Question 5

How can accessible regions be linked to the genes they regulate?

Accepted Answer

Nearest-gene assignment is convenient but frequently wrong, since regulatory contacts skip intervening genes. Stronger evidence comes from correlating accessibility with expression across many samples or cells, and from chromosome conformation data that reports physical contact directly. Any linkage claim should state which evidence type supports it.

Question 6

What changes when moving to single-cell accessibility profiling?

Accepted Answer

Each cell contributes very few insertions, so data are extremely sparse and binary in practice. Peak sets are usually defined on aggregated signal, dimensionality reduction replaces direct per-cell peak calling, and doublet detection becomes essential. Cluster stability should be tested across parameter settings before biological identity is assigned to any group.

Biotechnology

Biotechnology

Research Frontiers in Chromatin Accessibility ATAC-seq Analysis