How Can I Detect If A Protein Can Bind A Segment Of Dna Sequence In Vivo?

Question

[Ct586]

Hi, all,

I want to detect if a protein can bind single strand DNA sequence in vivo, which experimental methods should I use? And I do not know whether this protein has binding ability, is there any in silico method can predict this binding?

I prefer fluorescent label technique like FISH.

Thank you!

[Michael]

I would like to ask that question: why, given this question was posed on on a site dedicated to bioinformatics, do the first answers given feature the experimental part and not the computational prediction? Could it be that they have impression that the prediction tools have a terrible performance? From what I know, the accuracy of most the de-novo prediction methods is very low (this 2005 review states a correlation-coefficient of 0.2 http://www.ncbi.nlm.nih.gov/pubmed/15637633). Could somebody, link one or two more recent reviews to answer this?

[Neilfws]

This is a bioinformatics Q&A site, so we can help you with the in silico part - but ask elsewhere for experimental details.

[Casey Bergman]

Though I agree that this question does border on the edge of bioinformatics, I think it is still valid since it is about experimental design relevant to bioinformatics analysis. Similar questions about advice on experimental design/technology have been regarded as valid previously...e.g. Access Array (Platform From Fluidigm) For 454 Library Preparation Paired End Sequencing With Barcodes

[Michael]

Amazingly, the answers which you got up to now are about the experiments, not the in silico analysis.

[Neilfws]

Yes :) Can I ask that people restrain themselves, even when they know the answer, where questions are off-topic.

[Ct586]

I would not expect this question can evoke this debate. :)

Answer 1

[Chris]

Our group develops sequence-based structure/functions predictors. We also provide a DNA-binding predictor: DISIS

[Ct586]

Thank you! It is a great predictor.

Answer 2

[Casey Bergman]

Chromatin Immunoprecipitation followed by microarray analysis (ChIP-chip), high-throughput sequencing (ChIP-seq) or Real Time PCR (CHiP-PCR) are the industry standard for demonstrating protein-DNA interactions in vivo. All three techniques require in silico processing of the data. You can start your reading with the wikipedia entries on these topics:

• http://en.wikipedia.org/wiki/Chip-Sequencing
• http://en.wikipedia.org/wiki/ChIP-on-chip
• http://en.wikipedia.org/wiki/Real-time_polymerase_chain_reaction

[Ct586]

Thank you! Chip analysis is one of the best methods to detect this interaction. Also It supposed to reflect in vivo interaction. But are there any other methods which can do this, too? Such as FISH like http://www.piercenet.com/browse.cfm?fldID=B7A23C70-5056-8A76-4E39-B7DB24C3742C#fish?

Answer 3

[Asaf]

You could try the various servers for TFBS prediction but according to http://www.ncbi.nlm.nih.gov/pubmed/21304941 you won't get the answers you want unless you have DNase I digestion data.

[Ct586]

Great suggestions! Prediction is only the first step and only gives a concept that this protein may have the binding ability. Experimental results can give strong evidence.

Answer 4

[Alex Reynolds]

For motif binding instances, take a look at the FIMO tool, part of the MEME suite.

[Ct586]

Thank you! I will try to learn it.