DocumentCode :
769314
Title :
Limits of Genetic Control
Author :
Bremermann, H.J.
Author_Institution :
Department of Mathematics, University of California, Berkeley, Calif.
fYear :
1963
Firstpage :
200
Lastpage :
205
Abstract :
Control of behavior and specification of structure requires information. In simple cases lower bounds for the minimum amount of control information required can be computed. The amount of information in the genes of humans and related species is estimated (in different ways). By comparing the information available in the genes with the information required by certain structures and behavior patterns it can be decided what the genes can control and what not. Using an estimate derived from the number of nucleotides in DNA, it is shown that the genes cannot control the interconnections between individual neurons in the human brain in all potential complexity. They must be to some degree random or repetitive. Gene estimates derived from Drosophila studies result in lower estimates (104 to 3.5 ?? 105 bits). A stimulus-response pattern where all different possible response assignments are equally probable requires a minimum of n log2 n/e bits for n pairs. If antigen-antibody formation would satisfy this assumption, then the infonnation would well exceed the total information contained in the genes. Linguistic notions and their denotations may be considered as unconstrained stimulus-response pairs. A language of 4000 words requires about 5 ?? 104 bits. Hence a language of the complexity of human language cannot be innate.
Keywords :
Control systems; DNA; Genetics; Humans; Information theory; Military computing; Neurons; Quantum computing; Sequences; Upper bound;
fLanguage :
English
Journal_Title :
Military Electronics, IEEE Transactions on
Publisher :
ieee
ISSN :
0536-1559
Type :
jour
DOI :
10.1109/TME.1963.4323073
Filename :
4323073
Link To Document :
بازگشت