Are mutations actually random?

"Alfred Einstead" <[email hidden]> wrote in message
"]news:[email hidden]...

Quoted message said:

Looking at a DNA database, I notice that the patterns of
mutations, insertions and deletions are extremely non-
random. For instance, almost no insertions are of one type
(C, if I recall), whereas almost all deletions are of
another type (A, I think), and almost none involve G. By
far, most of all changes are of one type (between C and
T), and most of the remaining, by far, are of another type
(between A and G).

Is this accidental, or is it something that's
generally true?

One of the patterns you noted (C<->T and A<->G common, other
substitutions rare) is quite well known. Google for mutation
transition transversion

Your other observations are news to me, but perhaps not news
to those who are concerned with such things. Neo-Darwinism
does not assume that mutations are completely random, only
that they are random in their phenotypic effects. So, your
observations do not, by themselves, challenge the orthodoxy.

Alfred Einstead said:

Looking at a DNA database, I notice that the patterns of
mutations, insertions and deletions are extremely non-
random. For instance, almost no insertions are of one type
(C, if I recall), whereas almost all deletions are of
another type (A, I think), and almost none involve G. By
far, most of all changes are of one type (between C and
T), and most of the remaining, by far, are of another type
(between A and G).

Is this accidental, or is it something that's
generally true?

Check the accuracy of the sequences. Your insertion deletion
data has not been noticed by anyone else and they have been
looking for quite some time. At least I haven't heard of
them. There are machine errors in automated DNA base calling
that may cause certain types of miscalls. It used to be that
around half the time that you found a descrepancy in GenBank
sequences it was a sequencing error. I don't know what it is
now with people relying on machines for their data.

The biases noted so far have a basis in DNA chemistry.
Transition (pyrimidine to pyrimidine or purine to purine: T
to C, C to T, A to G, and G to A) mutations are much more
common than transversion (purine to pyrimidine like, T to A)
mutations. The repair and replication polymerases make
transition mistakes more often than transversion mistakes.
CpG sites have a very high frequency of mutation. In one
direction they are CpG and on the complimentary strand they
are CpG. Apparently the C changes to T on either strand at a
much higher rate than CpT or CpA sites. There are also
hotspots of mutation where we know that the mutations are
occurring at a very high rate, but we don't have any good
explanations. There is a base substitution that occurs at
the high rate of around 1 in 10,000 live births in the FGFR3
gene in humans. It causes dominant achondroplastic dwarfism
so we can get accurate counts. This is a CpG site, but it is
a much higher rate than a normal CpG site. This same
position isn't mutated at such a high rate in other animals
like mice or chickens. So mutations are not really random.
They are just unpredictable. We can determine that the site
in the FGFR3 gene in humans will change at a much higher
rate, but we can't tell you when or in what individuals the
mutation will occur in.

Ron Okimoto