combinatiorial control

spuriousmonkey

Banned
Banned
Dear Sciforumers,

as you might have read in another thread I recently was at a meeting on developmental biology and why not share my thoughts and experiences with you. After all, not everybody gets the chance to attend scientific meetings.

There was a very nice talk at this meeting by Susan Mango. She is quite a famous researcher in our field. She talked about several subjects, but one I found most interesting and that was 'combinatorial control'.

Combinatorial control deals with the problem of gene regulation. As you might realize it is essential that specific genes are switched on and off at specific moments during development. The expression of a gene at the wrong time or wrong place, or both can be disastrous for the developing embryo!

We all might have heard that a region of the DNA just before the actual gene regulates gene expression, and is called the ‘promotor’. On this promotor region DNA binding proteins can bind and actually determine if the gene is active or not. That is all old stuff.

And then came the idea of ‘control by combinatorial codes’ from the Drosophila field. The number of signalling pathways that actually regulate gene expression is tiny compared to the amount of different cell types. So how can these more or less general signals be translated into quite specific gene expression patterns? The idea of combinatorial codes is that gene expression regulation acts as some kind of computer where the quantity (amount of binding sites, amount of activators compared to repressors) and quality (strength of binding sites and activators and repressors) is sensed and calculated.

Short review: Nature (2000) vol 418, page 419-420

Susan Mango had another interesting twist on this idea that they actually had some algorithms looking for these short control elements in the DNA to which the regulators bind. They also noticed that by changing the DNA code of these elements slightly, for instance changing one nucleotide, the expression pattern of the target gene changed in time or location, or both.

Basically this is how evolution could manage to make chances fast. Genes do not change, but the promotor elements do, which has consequences for the temporal and spatial locations of certain genes, which can then influence shape and morphology.
 
Originally posted by spuriousmonkey
We all might have heard that a region of the DNA just before the actual gene regulates gene expression, and is called the ‘promotor’. On this promotor region DNA binding proteins can bind and actually determine if the gene is active or not. That is all old stuff.

And then came the idea of ‘control by combinatorial codes’ from the Drosophila field. The number of signalling pathways that actually regulate gene expression is tiny compared to the amount of different cell types. So how can these more or less general signals be translated into quite specific gene expression patterns? The idea of combinatorial codes is that gene expression regulation acts as some kind of computer where the quantity (amount of binding sites, amount of activators compared to repressors) and quality (strength of binding sites and activators and repressors) is sensed and calculated.

Short review: Nature (2000) vol 418, page 419-420

Susan Mango had another interesting twist on this idea that they actually had some algorithms looking for these short control elements in the DNA to which the regulators bind. They also noticed that by changing the DNA code of these elements slightly, for instance changing one nucleotide, the expression pattern of the target gene changed in time or location, or both.

Basically this is how evolution could manage to make chances fast. Genes do not change, but the promotor elements do, which has consequences for the temporal and spatial locations of certain genes, which can then influence shape and morphology.

I was just thinking the other day: Human DNA, and DNA in general, is incredibly compact. Just consider: 3 billion base pairs are equivalent to 6 billion bits, or 750 megabytes. This means that the program that creates the entire human organism can be stored on one computer CD. An installation of Microsoft Windows occupies nearly twice that amount of space. So either Microsoft Windows is more complex than a human body, or human DNA is algorithmically more efficient than Microsoft programmers.

Basically, what this means to me is that there must be a good number of nucleotides that when changed produce radical changes in the resulting organism. Think of DNA like an electronic file that has been ZIP'd. If you change one bit of a ZIP file, it is possible that the resulting unZIP'd file will be radically changed (much like DNA, it is also possible that the file will be currupted to the point where it cannot be unZIP'd, i.e., in terms of DNA, the change will end the development of the organism).

edit: bits not bytes!
 
Last edited:
developmental bio

The mechanisms organisms use to control gene expression are so diverse and complex that whole new levels of regulation are being explored that have nothing to do with DNA or mRNA expression. I'm a big fan of post-transcriptional gene regulation, but the field is comparitively in its infancy.
I have met Susan Mango and I agree she is a great scientist and very knowledgable. I bet the talk was good.
Promoter bashing has revealed a great deal about the sequences you speak of, but there is far more to know.
 
Re: developmental bio

Originally posted by wrmgrl
The mechanisms organisms use to control gene expression are so diverse and complex that whole new levels of regulation are being explored that have nothing to do with DNA or mRNA expression. I'm a big fan of post-transcriptional gene regulation, but the field is comparitively in its infancy.
I have met Susan Mango and I agree she is a great scientist and very knowledgable. I bet the talk was good.
Promoter bashing has revealed a great deal about the sequences you speak of, but there is far more to know.

Is there any indication that some part of the mechanism for gene regulation is itself not the direct result of gene expression? In other words, is there any indication that there exists some vital part of an organism that is not itself encoded in the DNA, so that even if we knew exactly how to interpret DNA in its entirety, we would still be missing something if we did not have at least one specimen of a living organism containing that DNA?
 
Re: Re: developmental bio

Originally posted by nathan_w_cheng
Is there any indication that some part of the mechanism for gene regulation is itself not the direct result of gene expression? In other words, is there any indication that there exists some vital part of an organism that is not itself encoded in the DNA, so that even if we knew exactly how to interpret DNA in its entirety, we would still be missing something if we did not have at least one specimen of a living organism containing that DNA?
I suppose the answer to that question is both yes and no. It's called epigenetics, the process by which gene expression is altered without changes in the gene or its regulatory sequences. Mainly the no part comes down to the technical definition of a gene. There are non-translatable RNAs produced that actually regulate the expression of mRNAs. In other words, RNAs can regulate protein expression without every producing their own protein. Theoretically, once they understand all the mechanisms by which genes can be regulated, transcriptionally and translationally, we would not require living organisms to understand how gene expression is regulated. With the complexity of these pathway, I think this understanding is not likely to be soon achieved. There are no protein or RNA determinants that are not encoded in the genome, it's just they do not necessarily exist in the form of genes.
 
Re: Re: Re: developmental bio

Originally posted by wrmgrl
I suppose the answer to that question is both yes and no. It's called epigenetics, the process by which gene expression is altered without changes in the gene or its regulatory sequences. Mainly the no part comes down to the technical definition of a gene. There are non-translatable RNAs produced that actually regulate the expression of mRNAs. In other words, RNAs can regulate protein expression without every producing their own protein. Theoretically, once they understand all the mechanisms by which genes can be regulated, transcriptionally and translationally, we would not require living organisms to understand how gene expression is regulated. With the complexity of these pathway, I think this understanding is not likely to be soon achieved. There are no protein or RNA determinants that are not encoded in the genome, it's just they do not necessarily exist in the form of genes.

Interesting. I think I actually understand what you mean.
 
Re: Re: Re: Re: developmental bio

Originally posted by nathan_w_cheng
Interesting. I think I actually understand what you mean.
I am glad to hear it. I am working on my comprehensive exam so all this stuff is fresh in my head. But I'm also low on sleep so I was afraid I'd be incomprehensible.
 
I am aware that there are others level of control.

But for morphogenesis this mechanism is particular attractive, especially in the light of the new views.
 
Originally posted by spuriousmonkey
The idea of combinatorial codes is that gene expression regulation acts as some kind of computer where the quantity (amount of binding sites, amount of activators compared to repressors) and quality (strength of binding sites and activators and repressors) is sensed and calculated.

Was there any discussion as to the chemical or perhaps purely physical (e.g. by electric charge) mechanism by which this computation takes place and results in gene expression?
 
Back
Top