J. David Spafford wrote:
>> Hi. I am looking for some advice. I have cloned a voltage-gated sodium
> channel gene in both its cDNA form and the full-length region covered in
> genomic DNA. I have found that the majority of the introns are conserved in
> my gene and in all genes of the family. The major differences arises in the
> unconserved non-transmembranal regions of the protein. If I do a comparison
> with the sodium channel with its known ancestor, the calcium channel, I am
^^^^^^^^
you mean "relative"?
> also finding some conservation of introns. However, there appears to be
> intron slippage, because some of the introns aren't in the exact homologous
> locations. In some cases, a poor amino-acid alignment may be the cause of
> the intron slippage. Or maybe, I am using my a priori knowledge about
> kinship and trying to see a pattern of conservation that isn't present in
> the data. I have difficulty deciding whether I should consider one to ten
> bp slippage of an intron location warrant defining an intron as homologous
> orb non-homologous.
This is the right sort of question to be asking. Its possible that
introns
"slip" or "slide", but there isn't any direct evidence that this
happens,
or even any indirect evidence that it is common. Most of the "evidence"
is just the sort of pattern that you are describing-- some introns match
exactly in position, while others do not. Whether such patterns
indicate
a special phenomenon of "sliding" is unclear. There are now many
examples
where such discordant introns are due to sequence errors and alignment
ambiguities (we are about to publish a list of 20 of them). If the
close-but-non-matching intron positions *have the same phase*, then you
may be looking at the effects of alignment ambiguities.
If the sequences and alignments seem solid, then you could be looking at
a real biological phenomenon-- either due to "sliding", or to separate
events of loss and gain of introns. In distinguishing these it will be
helpful to look at the distances between intron positions, as well as
the
phylogenetic distribution of the intron positions. The phylogeny may
suggest that an ancestral intron hopped from to another position
(usually,
the phylogeny does not suggest this, but instead suggests that the two
introns
were separately gained). With respect to the distances between introns,
it is not usually clear that the non-matching intron positions are
really
so close as to require some special explanation based on "sliding". The
"birthday problem" is relevant here-- the chance that a room with X
people
will have two people with a matching birthday exceeds 50% for X >= 23.
By analogy, if we have X uniformly IID introns, each added to a separate
copy of a gene with 365 codons (i.e., same as the numbers of days
per year), the chance that two will be *in the same codon* exceeds 50%
for
X >= 23. Actual data are more difficult to analyze than this simple
example
would suggest, but you get the idea.
Its important to realize that you probably cannot get solid answers for
a single (small) set of data about whether or not "sliding"
is responsible for differences. If "sliding" occurs at all, it is
probably
rather rare. How many different genes and intron
positions are included in your set of channel gene data?
Arlin
--
Arlin Stoltzfus, Ph.D. (arlin at is.dal.ca)
Department of Biochemistry, Dalhousie University
Halifax, Nova Scotia B3H 4H7 CANADA
phone: 902-494-2968 fax: 902-494-1355
