some questions about ChIP-Seq replicates/controls

[hongyan]

1. Design: Does ChIP-seq need biological replicates? Is control sample necessary?

2. DNA shearing: If doing ChIP-chip, DNA fragments usually range from 500-600 bp. But when illumina make libraries, they start from DNA with ~200 bp . Do we need to shear DNA to ~200bp during ChIP?

[Chipper]

1. No. No. But of course multiple replicates and controls can give you more reliable data.

2. It is possible even if the majority of reads are above the selected size, or to select larger fragments. But around 200 bp seems to be optimal for Illumina and SOLiD to get good number of reads and high resolution.

[hongyan]

Hi, Chipper,
Thanks for your reply.
If my budget is enough for only two samples, which one should I choose: two biological replicates or one sample/ one control?

As I understand, increasing the depth of sequencing enough for one sample may have the similar power as doing two biological replicates. Is it right?

[Chipper]

One sample and one control or two samples. You can get god results from one lane and no control, but have to be aware that some regions may act strangely due to duplications etc. You can re-use the library to increase sequence depth later on, it should be cheaper than running two different libraries.