Hariri A, Weber B, Olmsted J
Department of Chemistry and Biochemistry, California State University, Fullerton 92634.
J Theor Biol. 1990 Nov 21;147(2):235-54. doi: 10.1016/s0022-5193(05)80054-2.
The usefulness of information-theoretic measures of the Shannon-Weaver type, when applied to molecular biological systems such as DNA or protein sequences, has been critically evaluated. It is shown that entropy can be re-expressed in dimensionless terms, thereby making it commensurate with information. Further, we have identified processes in which entropy S and information H change in opposite directions. These processes of opposing signs for delta S and delta H demonstrate that while the Second Law of Thermodynamics mandates that entropy always increases, it places no such restrictions on changes in information. Additionally, we have developed equations permitting information calculations, incorporating conditional occurrence probabilities, on DNA and protein sequences. When the results of such calculations are compared for sequences of various general types, there are no informational content patterns. We conclude that information-theoretic calculations of the present level of sophistication do not provide any useful insights into molecular biological sequences.
对香农 - 韦弗类型的信息论测度应用于诸如DNA或蛋白质序列等分子生物学系统时的有用性进行了严格评估。结果表明,熵可以用无量纲的形式重新表示,从而使其与信息具有可比性。此外,我们已经确定了熵S和信息H朝着相反方向变化的过程。这些ΔS和ΔH符号相反的过程表明,虽然热力学第二定律要求熵总是增加,但对信息的变化并没有这样的限制。此外,我们已经开发出了一些方程,用于在DNA和蛋白质序列上结合条件出现概率进行信息计算。当对各种一般类型的序列进行此类计算的结果进行比较时,不存在信息内容模式。我们得出结论,目前复杂程度的信息论计算并不能为分子生物学序列提供任何有用的见解。