A Molecular LAN Oct 94

This idea stems from the emerging techniques in Molecular Engineering (MOLE), arising in several disciplines. I refer to such things as Directed Molecular Evolution, Atomic Force Microscopy, DNA Sequencing, Micro Machining, Molecular Modelling.

I envisage a Molecular Engineer (several years in the future) sitting at a MOLE workstation designing a molecule. He will have access, via a LAN, to a molecular database (like the small organic mole Database at Cambridge). He can call up molecules from this and manipulate them and apply various software tests and transformations. Eventually though, the software has to be transformed into hardware, so he will have to call for one or more source molecules from a molecule store (rather like the RFLP stores for genome sequencing). He can then use these molecules, perhaps as seeds for Molecular Evolution, or within some atomic manipulation device for manual modification. After construction, the molecules will then need transporting on to test, QA, and finally to the molecule store. The details of the devices used do not concern me here; what I am trying to establish is the need for the selection and transport of quantities of molecules amongst and between a number of distributed MOLE stations, all by remote control. The requirement is very similar to a LAN (Local Area Network) for moving data between workstations, and it was from this analogy that I instantly saw a possible solution.

All MOLE stations are connected by a MOLAN, along with the normal LAN. The physical connection is via a ring of flexible tubing which contains an inert carrier (fluid or gas). This carrier is pumped continuously around the ring. Each station on the MOLAN is connected to the tubing, and can potentially act as transmitter or receiver of molecules. Now comes the clever bit. Each station has a unique address, consisting of a genetic base sequence (i.e. of CGAT). To transmit a molecule, it is tagged with a base sequence of the target station, along with an address packet header and trailer, and then injected into the tubing. Each station also has a receiver, through which the carrier passes, with molecular 'traps' keyed to its own bases sequence address. Thus when a molecule with the correct address comes along, it is captured and then transferred to the stations 'input port', after cutting off the address packet. This is very similar to a token ring LAN for data packets.

If molecules are not trapped first time, they may pass around the network and be trapped on another pass. Probably there will need to be some garbage collection mechanism, which could be as simple as a periodic flush out.

Of course, I have omitted many details - how long does the base sequence need to be, how much redundancy is needed, what sort of header/trailer is needed, how are the captured molecules removed etc. These details will need to be resolved in R & D of the MOLAN; I do not under estimate the difficulties, but neither do I see them as insurmountable. MOLAN stations such as I describe will not be cheap, but then MOLE has huge potential pay back. Many of the techniques I describe are already being well used in genetic engineering, and their cost will undoubtedly fall with increasing volume as MOLE grows in importance. As often, I am probably several years ahead of the market with this idea, but that just gives you longer to develop it.