One of the quaint notions creationists toss out occasionally is that evolution can only destroy information
in DNA, not increase it. While we have already proved that false with
various experiments, I think it stems to a bad understanding of what
mutations are.
One of the most common mutations is sequence duplications. A section of DNA gets copied several times by mistake (one of the reasons we have 3bn base-pairs in the human genome but only use 23,000 genes to code for a person). While intuitively, it might seem as if these copies don’t add any new information (they have the same codes as the original), these new sequences do create strings of new base-pairs that can yield new functions. (In information-theory, this is equivalent to an increase in Kolmogorov complexity).
To make this more intuitive, consider the following little experiment. I’ve generated a column of random letters and found a set of words in there. So, I’m using an lexicographic algorithm as an analogy for a genetic. Lexicographic codes follow rules too- they have to be read from left-to-right they must contain a vowel, each letter can only be used once.
The second column is now the duplicate, and I’ve highlighted now in green, the new words this
duplicate allows me to form. Some of the new words in the 2nd column
are completely new- e.g. smug, win, are. Simply by duplicating strings
(a mutation) of old information, we can generate entirely new information as code.*
* Disclaimer- this is an argument form analogy and is not representative of how genetic algorithms treat sequences.
One of the most common mutations is sequence duplications. A section of DNA gets copied several times by mistake (one of the reasons we have 3bn base-pairs in the human genome but only use 23,000 genes to code for a person). While intuitively, it might seem as if these copies don’t add any new information (they have the same codes as the original), these new sequences do create strings of new base-pairs that can yield new functions. (In information-theory, this is equivalent to an increase in Kolmogorov complexity).
To make this more intuitive, consider the following little experiment. I’ve generated a column of random letters and found a set of words in there. So, I’m using an lexicographic algorithm as an analogy for a genetic. Lexicographic codes follow rules too- they have to be read from left-to-right they must contain a vowel, each letter can only be used once.
The second column is now the duplicate, and I’ve highlighted now in green, the new words this
duplicate allows me to form. Some of the new words in the 2nd column
are completely new- e.g. smug, win, are. Simply by duplicating strings
(a mutation) of old information, we can generate entirely new information as code.** Disclaimer- this is an argument form analogy and is not representative of how genetic algorithms treat sequences.
No comments:
Post a Comment