Linguistics And Biological Chemistry

Judith Klein-Seetharaman and Raj Reddy, Carnegie Mellon University

How can we improve the nation's productivity and quality of life in the next 10-20 years? The nation's performance is dependent on functions of the human body, since they directly or indirectly determine human ability to perform various tasks. There are two types of human ability: (1) "inherent abilities," tasks that humans are able to perform, and (2) "external abilities," tasks that we cannot perform per se, but for which we can design machines to perform them. Both categories have individually experienced groundbreaking advances during the last decade. Our inherent abilities in terms of fighting diseases, repair of malfunctioning organs through artificial implants, and increased longevity have greatly improved thanks to advances in the medical and life sciences. Similarly, technology has provided us with remarkable tools such as smaller and more efficient computers; the Internet; and safer, cleaner, and cheaper means of transport.

Integration of Inherent and External Human Abilities

Advancing our society further necessitates a better integration between the inherent and external abilities. For example, interfacing computers with humans need not require keyboard and mouse: ongoing efforts advance utilization of speech interfaces. But ultimately, it would be desirable to directly interface with the human brain and other organs. This will require further advances in elucidating the fundamental biological mechanisms through which humans think, memorize, sense, communicate, and act. Understanding these mechanisms will allow us to (a) modify our inherent abilities where natural evolution does not feel any pressure for improvement and (b) design interfaces that connect our inherent abilities with external abilities.

Grand Challenge: Mapping Genome Sequence Instructions to Inherent Abilities

How can we aim to understand complex biological systems at a level of detail sufficient to improve upon them and build interfaces to external machines? In principle, all the information to build complex biological systems is stored in an "instruction manual," an organism's (e.g., a human's) genome. While we have recently witnessed the elucidation of the entire human genome sequence, the next logical grand challenge for the coming decade is to map the genome sequence information to biological functions. Interfacing between biological functions and artificially manufactured devices will require improved structure-property understanding as well as manufacturability at a multiscale level ranging from Angstrom-sized individual components of biological molecules to macroscopic responses. This will be possible through existing and future advances in nanotechnology, biological sciences, information technology and cognitive sciences (NBIC).


The sequence ^ function mapping question is conceptually similar to the mapping of words to meaning in linguistics (Figure F.6). This suggests an opportunity to converge two technologies to address this challenge: computational linguistics and biological chemistry, via "biological language modeling." The term "biological chemistry" is used here to stand for interdisciplinary studies of biological systems, including biochemistry, molecular biology, structural biology, biophysics, genetics, pharmacology, biomedicine, biotechnology, genomics, and proteomics. The specific convergence of linguistics and biological chemistry is described below under the heading, "The Role of Converging Technologies: Computational Linguistics and Biological Chemistry." Its relation to the more general convergence with NBIC is described in section, "The Role of Converging Technologies: NBIC and Biological Language Modeling." Two specific applications of linguistic analysis to biological sequences are given in "The Transforming Strategy," to demonstrate the transforming strategy by example. If we can solve the sequence ^ function mapping question, the implications for human performance and productivity are essentially unlimited. We have chosen a few practical examples to illustrate the scope of possibilities ("The Estimated Implications"). Implications for society are sketched in "Implications for Society," followed by a brief summary.

Biology: Language:

Figure F.6. Analogy between language and biology, which forms the basis for the convergence of computational linguistics and biological chemistry.
0 0

Post a comment