Guangnan Tian June 2
Last week's meeting with Yu Tao did give me some concern that without a concrete model, the team is difficult to run efficiently only to reject other people's preliminary ideas.
I took a brief look at several papers from Elowitz and Collins on Noise, Stochastic study and Oscillators and tried to find some ideas on a concrete model, which can takes noise into consideration. And I do get a model.
Since I saw some teams using fusion promoter/operator sequence, so that it is possible to buid two identical promoters of different orientations, which partially overlap ( ------- ________ ------- ________ )for example, so that the first transcription event is mutually exclusive. It is like a toggle switch basically, but mechanisms were quite different, but the kind of toggle switch inducer can works as reseting.
In terms of intrinsic noise and/or extrinsic noise, it is actually not a problem for us, since for the Elowits paper, he integrate the two promoters about 1.5 mb from the ori, which means that they are seperated by half genome size, and that is actually quite large difference of intracelluar local environment, and it is the repression system which cause a 20 fold decrease of expression base level that increase the so called Noise in some strains.
I am thinking about the two promoters can be modified, adding downstream operating sequence, that can mutually repress each other (which must be quite tight), and since two promoters are adjacent, the repression can be quite efficient, In terms of the extrinsic noise, since the expression level in different cells can vary. I am planning not to incorporate reporter genes directly into the cistron, rather make the two promoters coding for repressors for two reporter genes, then the one who gives reporter signals is actually a constituitive promoter controled genes, and the noise is rather small, the cell cell difference can be reduced to as little as 10%
Xiuye did mentioned that this project mean lack certain big " usage" or " meaning" by itself. It is actually important to make the readout machine.
I have talked with Xiuye, and tried to find some researches on counting machines, but yet I haven't find any natual processes which can lead to quantuminzed counting.
But I do have a new idea about what to do next
It is actually a memmorizer which can forget and select information.
Since you told me that some bacteria which can remember turned to fail to forget.
Since what about we give up the previous model that let one cell has 3 circuits of randomizer, but we have a circuilt, which can remember the state of the randomizer in randomizing event 1,
then the machine can reset the randomizer and remember the second outcom of randomizer as state 2.
If state1 = state2 then output A If state1 <> state 2 then output B
And the machine can reset the randomizer again, and the third outcom of randomizer as state 3.
If state2= state 3 then output A If state2<>state 3 then output B
This can go on, so the memorizer can compare the present state with the last state.
For this machine is will be much easier, Reference is the ETH project and Tianjin project in 2007.
I will try to build up a whole model for the two circuits, and try to make it concrete to parts level.
Since I am busy with tests these days, I am not very sure if I could make it before June 2nd, but I will try my best.
Xiuye and I also talked about possibilities of proposing different ideas or different subject. While time is really limited, from my observation, previous teams actually build quite simple circuits, and we don't need to make it very difficult.
There is also an interesting paper I wish to share, when I read Elowits paper I got it by chance. It is propagation noise, the noise given rise by sequential activation/repression of downstream factors.
ur name date