Differnces between PWM derived from SELEX and ChIP-seq

Question

[michaelchen33]

As title.

Because SELEX experiment can detect strong binding site of TF, and ChIP-seq experiment can detect both strong and weak binding site (due to the cross-linked).

So I wonder if PWM derived from SELEX still have the strong binding site property and PWM derived from ChIP-seq still have the strong and weak binding site property.

Or,it will lose the property when these two experiments model to PWM.

Any suggestion or papers will help.

Thanks

Answer 1

[xb]

The TFBSs by ChIP-seq data are usually divided into different groups (strong and weak, sometimes with an intermediate group: moderate) based on their binding strength. A common assumption is that "ChIP-seq signal values are correlated with the affinity of the TF-DNA binding" [ * ]

PWM alone helps us model the sequences at TFBSs but no such information about binding affinity, unless binding scores are also annotated.

[michaelchen33]

So does it means that when I transformed the experiment data like ChIP-seq to PWM, I lost the binding affinity information even through I use the strong group of ChIP-seq data?

Thanks

[xb]

Yes, if you have the matrix only.