If you watch television news these days, you can spot bias a mile away. Well, we can’t do much about shoddy journalism, but a new report from scientists at UC Berkeley, led by Michael Eisen, calls attention to some bias we can fix; the kind caused by the structure of chromatin in ChIP experiments.
The Eisen group analyzed comparisons between data generated from ChIP (cross-linked) and regular genomic input DNA samples. They uncovered many biases, such as:
- Under/over-representation of specific chromosomal regions in deep sequencing
- Over-representation of transcriptionally-active regions in deep sequencing
- Differential sized fragment shearing in distinct genomic loci
Their work shows that different chromatin structures, preserved through cross-linking, can have downstream effects on both DNA shearing and stability, ultimately leading to biases into any downstream quantification or detection. They go on to suggest that these biases could in turn be capitalized upon to detect specific chromatin structures in future studies.
To get a full dose of chromatin bias check out PLoS One, August 2009.