7.4 Biocomputers

The integration of various electronic devices has led to the invention of the presentday computer. In this sense, if we can integrate biomolecular electronic devices as we wish, a new computer will be developed. We can call it a biocomputer. But what will a biocomputer look like? How will it work, and what will it do? Comparison between a brain and a computer will help us illustrate. You can easily recognize your wife or husband, your daughter or son among many people walking in the airport or a station. You can also easily imagine what kind of present will be good for your parents. Our brain is very good at recognition and imagination. Computers cannot defeat us in these respects. On the other hand, you will easily forget the name or the face of the person you met at a business party. When it comes to memorization, we are very poor compared to a computer. A biocomputer will be, and should be, one that can function like our brain. If we define a biocomputer that is very good at a certain performance (like recognition of images, of sounds, or imagination) as a single-function biocomputer (SFB), future biocomputers will be used by connecting SFBs in a web. Biomaterials such as enzymes or antibodies are not as stable as silicon-based LSI. We have started work on synthesizing a 100 percent artificial enzyme or antibody (Rachkov et al. 2000). In this way, a biocomputer could be used without "getting tired". Up to now, artificial antibodies have much lower susceptibilities and activities than natural antibodies. Natural antibodies/elements are extremely flexible, which is thought to contribute to their good discrimination. If more flexible (but long-lived) artificial antibodies and elements can be created, this should result in a striking improvement in materials made from them.

Such biocomputers will play an important role, for example, in an aging society. If a machine can recognize what it saw or heard, it will help the development of a nursing robot. Soon we will have the whole information on our genome, and part of our tendencies toward diseases will be predicted by the result of genetic diagnosis. Biocomputers will be a good interface between patients and the medical care system. Furthermore, the development of a biocomputer itself is an act of elucidation of our brain. It is, in a sense, the act of understanding ourselves.

Recently, a trial was carried out to solve a mathematical problem using DNA molecules (Adleman 1994). In this work, the high selectivity of DNA molecules was used for calculation. In a method of calculation using DNA, one parameter is addressed to a DNA molecule of a certain thirty-base sequence (Sakamoto et al., 2000). A DNA sequence that is complemental to the other sequence was defined as a negation of the parameter expressed with the counterpart DNA sequence. These two parts, the original and its counterpart, would make a complemental binding and make a hairpin structure. DNAs that have hairpin structures can be regarded as "false" solutions. In such a way, DNA computation is performed. This is just a very beginning of the molecular computation, and its future image is now just becoming manifest.