ChipSeq multiplexing the number of TF

Question

[grant.hovhannisyan]

Hi Biostars,

Sorry for my ignorant question, but I couldn't find the answer in internet.

In one ChipSeq experiment, how many transcription factors can be analyzed?

Thanks

Answer 1

[Devon Ryan]

As many as you want. You're only going to IP one thing at a time, but how many you multiplex together is up to you.

[grant.hovhannisyan]

Thank you Devon! This is actually where I'm a bit confused: of course perfoming multiplexed sequencing is not a problem, but is it practically possible and feasible (time, costs) to IP all TFs in S. cerevisiae (ca 300), for example?

[GenoMax]

I suppose anything is possible, if you throw enough resources at it. Question is what resources do you have access to.

[grant.hovhannisyan]

Makes sense. I have no clue about costs per TF, but I wouldn't spend more than 5K EUR for this experiment.

[Devon Ryan]

For fruit fly ChIPseq people tend to use ~20 million reads per sample/ChIP, so given a genome 1/10th of the size you might get away with ~2 million reads per sample/ChIP. That's making a lot of assumptions that I'm not sure are actually valid. I would encourage you to find a public yeast TF IP dataset or two and try to guestimate from that how many reads you can get away with. Then you can more reliably estimate the actual costs. Please actual get more sequencing than you actually need, since inevitably there will be a bit of unequal pooling.

[GenoMax]

Start with the cost of sequencing and subtract that from 5K EUR. Then work backward to see how many TF samples you can fit in in the number of reads you expect to get (unless you were not counting sequencing cost in 5K).

[grant.hovhannisyan]

Thanks for suggestion, but without knowing the cost of IP per TF would be hard to plan anything. On other hand it seems that even for S. cerevisiae antibodies for different TF are quite limited, so analyzing the whole TFome would be very tough.