"

From 2012.igem.org

== '''Overall project - Associative Memory Network Using Bacteria ''' == Since we are a big team and it's our first participation on iGEM, we have two projects going on at the same time. The first one is "Associative Memory Network Using Bacteria" and "Plug&Play Plasmid" === Associative Memory Network Using Bacteria === The memory storage in biological systems has a critical role in biotechnology development. A systemic way of storing a specific memory that can be recovered and used at any moment is studied in several experiments and mathematical models involving neural networks. One of these models, known by “Hopfield Network”, considers the memory storage as a neurons association that shares a characteristic pattern of “communication intensity” – the “measure unity” of a neuron network . This model is notorious for allowing systems the recognizing of patterns. In this project we propose the genesis of aa communication network using E.coli populations, in order to create a system with associative memory ,like a Hopfield Model. Genetically different populations will be generated and isolated from each other, keeping contact by means of Quorum Sensing Substances (QSS). These QSS will be responsible by inhibition or excitation of pre-determined populations, measuring the amount of excitation by means of GFP fluorescent level. The final objective is to achieve a specific complete pattern of excited and inhibited populations by means of the interactions between bacteria populations, based on an given incomplete pattern. The network will behave in way of recognizing this pattern and choose between two systemic memories inserted on the communication network. == Project Details== === Physical System === In a neuronal communication system, the cells have specific locations and the information is sent by a direct physical connection(the axon of the neuron). An important aspect of the specific location of the neurons is the chance to observe the neuronal activation, which occurs always in the same area. This process in a bacteria network is more complex than that, because there is no communication between only two cells in a specific point of the space, but between various populations of bacterias that have movement , in other words, there is no specific point in space where you can observe the same phenomenon in different moments. To replicate the dinamic of a physical neuron network in a moving cells system,it would be necessary to stablish a way to give a specific adressing to the exchange of informations between the components of the system.[[File:Physicalsystemforbacterialnetwork.png‎|thumb|upright=1.0|]] In order to give an specific addressing to the bacterial population, a system will be constructed in a way to make 9 different E. coli populations communicate with each other and keep their positions. The system allows communications not only between neighbors, but integrating signals coming from the hole system, activating or inhibiting the answer of the cell according to the specific memory established. A model of the system is showed in the figure ->. This system has 96 cavities with permeable membranes that allows the exchange of signals by "quorun sense" substances, but keeps the populations in a specific location. In other words, it allows the communication, but not the mix of different populations, keeping the signals uniform according to their origins. === Genetical Construction === On the registry of parts there is only 4 well characterized quorun sense systems, these are insuficient to the 9 different quorun sense substances required initially. So, as a proof of concept, we took 2 particular conformations of memories that use only two quorun sense substances. This conformations ("X" and "0") don't have points of intersection, what means that all activated positions in "X" inhibit all activated positions in "0". Consequently, all activated positions in 0 inhibit all activated positions in X. By this way, there isn't need of distinction between a huge variety of signals, and, consequently, the positions of each "pattern" can be programed with the same genetic code, utilizing two sistems of quorun sense to discern between two different inputs(incomplete patterns). The two systems of quorun sense were choosed because the difference between their substances, minimizing the probability of misreading and enhancing the effects. In order to convert the signal in activation or inhibition,there will be a "competition" between the quorun sense inhibitors and quorun sense activators for the transcription sites in a population. In other words, if there is more activators for "X" pattern than for "0" pattern in a position, then the "X" will be favored, occuring the same clash in the other positions according to the input signal(incomplete pattern that will activate memory).[[File:Systemtest.png|thumb|upright=1.0|]] The multi regulated promoter is the Prm, whose inhibitor is the transcption factor cl434 and activator is the cl. The genetic design for "X" and "0" can be seen in the figures. The use of cl1434 and cl in the signal transduction allows the generation of associative memory systems limited only by the variety of quorun sense systens available. === The Experiments === === Part 3 === == Results ==