Confused about ChIP-chip - Please help with answer, I suddenly feel my results are really dull (Jan/28/2009 )

Hello everyone!


I was a bit puzzled today when someone in my new lab asked me about what I had done in my thesis and I told her I had done ChIP-chip. I told her it had been great because I had had significant results and fold regulations. She then asked me if ChIP-chip results normally correlated with expression arrays. I answered that as far as I knew, they normally didnt and that my old boss had been asked about it once and he had responded that it didnt but he believed that it was a useful method to check out the possible underlying patterns that dictate expression of other genes indirectly.


Im not even a PhD student yet and she is almost a Dr. so I was a bit short of answers for all her questions. She proceeded to ask me then why H3K4 methylation or H3K9 methylation were activating or silencing marks if their deregulation in a ChIP chip experiment wasnt neccesarily reflected in an expression array too.
She didnt understand what it all meant. You have a gene list for a ChIP-chip against me3H3K4. Results are significant, ratios are awesome but what does it tell you? if anything at all? I told her you look at correlations with other marks, you look at gene ontology, you look at locations blablabla and she was like yeah you get more and more lists of genes but what do u do functionaly? How do you sell your story? One gene promoter is more acetylated in a cancer cell and in a normal cell yet the expression is the same. Great! So why do I care?

This isnt the first time that someone has asked me such questions. A Professor also asked me what it all mean if I didnt know how it was all happening and what it meant but this time I really felt quite stupid.

Help!!

PhD on Jan 28 2009, 02:41 PM said:

PhD on Jan 28 2009, 02:41 PM said:

I'm guessing it was an undergraduate thesis, master's thesis?

PhD on Jan 28 2009, 02:41 PM said:

in ChIP-ChIP, you are looking for where your protein of interest is binding. you pull down your protein that is crosslinked to DNA, then release the protein, and all the DNA is put on a microarray chip that has probes for genomic-wide sequences. more commonly, the microarray chips will cover only promoter and orf regions since thats where most regulatory elements will be found. in contrast to ChIP-chip which is specific for the protein of interest, expression arrays probe for all expressed RNAs. Thus, if a protein binds to a specific region on the genome (say a promoter) it may be likely that it correlates to increase in mRNA = increase in expression. But that is not always the case.

PhD on Jan 28 2009, 02:41 PM said:

That question is a bit tricky because it has been proposed that histone methylation is correlated with heterochromatin. Since the chromatin is tightly packed, it would be difficult to create the transcription bubble necessary for transcription, and therefore help to silence genes. histone methylation has been known to help recruit additional silencing mechanism on the DNA level. if these methylation marks have been de-regulated and there is no change in expression, it may suggest that additional activating factors are required for transcription. however, the trend is that methylation is associated with gene silencing on a genome wide level. because it is genome-wide analysis, you'd expect exceptions and that is why you take ratios to determine the most LIKELY and most SIGNIFICANT of the entire dataset. i could go into more detail, but i wont bore you. of course, i've never done a chip-chip myself, so take from that what you will.

PhD on Jan 28 2009, 02:41 PM said:

you shouldn't care, the acetylation mark is useless here for this one gene. if you are deregulation a HDAC or HAT, then it will be more interesting to look at genes that are affected. since gene regulation is a highly complex system, the acetylation mark at that locus are lacking transcription activating factors. other mechanisms are possible, but the overall idea is that in cancer multiple events must happen, and for this locus, the acetylation was not sufficient to change expression. however, since it is a cancer cell, there must be other genes that have been affected by other epigenetic marks, and the challenge now is to determine what marks and which genes and its significance.

PhD on Jan 28 2009, 02:41 PM said:

not to be cruel, but please do get a better idea of why you are doing each experiment, how you interpret your results, and the shortcomings of your method. i understand that you do not have a phd, but when you do become a phd student, please remember to keep these in mind. more importantly, however, i feel that even if you try to wholly understand your experiment, in science there will always be critical people and you will occasionally be stumped, but dont be afraid to admit that you dont know. its much better than spewing balony.

Thank you so much for that answer!
It was indeed an undergraduate thesis and dont worry you havent been cruel at all you just confirmed all the things that I had been thinking but never dared to say because my exboss was so excited about the patterns and fold ratios that we saw that I figured in the end they would perform extra experiments to round up the story after I left. Also when he was asked the same question in a congress the answer was very short and kind of uncomfortable and people werent truelly convinced.

It was the answer I feared hehe I had thought about this during the thesis but never dared to flat out point out that it was a weak story and that the project maybe hadnt been planned that well. I have a great relationship with my ex supersivor, she is a great person and has always been very supportive. I thought that what I was thinking was arrogant and who would attack my results like that nways? Of course I was wrong because in my thesis defense I had to answer these questions and I had but a weak argumentation for the main purpose of the project because it was in fact extremely leaky. Oh well.... I am still glad that ive learned all this and im glad this girl in lab asked me things in a way that I had to finally give up and agree with her because I did hehe.



Have a nice day!!