Use of secondary structure in alignments

Brian Foley brianf at
Thu May 23 12:01:23 EST 1996

Diane Stothard (dstothar at wrote:

: I was wondering what the consensus feeling is in this group 
: for using secondary structure as a guide for aligning rRNA 
: gene sequences.
: ...

	I wouldn't profess to know "the consensus of the group"
but here are my thoughts:
	Some ribosomal RNA base-paired structures are known
with a high degree of certainty (i.e. not just predicted by
reverse-complementarity, but also shown to be conserved
through evolution, or know to be needed for function via 
mutagenesis studies).  Mutations within these regions face
different selective presure than unpaired positions.  If
a "G" in a "G-C" base-pair mutates to "A" and the organism
can survive, future mutations which either revert the 
"A" back to "G" or convert the "C" to a "T" to restore 
base-pairing, will be selected for.
	If we are interested strictly in molecular evolution
to determine phylogeny, we want to know only what DNA
sequence changes actually ocurred.  And base-paired
regions may slide along a sequence, or jump to new
locations, thus the new base paired region has
functional homology to the old, but is evolutionarily
derived from a diferent region of DNA.  These 
jumps and shifts are rare, compared to single base
changes within each side of a base-paired stem, but
they do occur.  If a new region of DNA is now base-
pairing to form a stem with the old region, the
new region will likely have a similar sequence,
and the old region which no longer base-pairs will
be free of selection toward base-pairing.  Thus 
aligning the new stem structures to the old, will be 
misleading for phylogenetic studies.
	If we are interested in the function of ribosomal
RNA, or interested in how large evolutionary leaps are
made (a stem structure moving in the rRNA, rather than
point mutations), then we need to look at secondary
structures, and aligning such structures may be very 
useful, even if they are not derived from the
same regions of DNA.
	The Human Immunodeficiency Virus and other 
lentiviruses have some important secondary structure
regions involved in either the ss-RNA portion of their
life cycle, or in regulation of mRNA production during the
ds-DNA proviral phase.  These viruses evolve extremely
rapidly, thus giving us a chance to sample significant 
evolutionary changes over time.  Perhaps the intensive
sequencing efforts on these viruses can provide some
insight into evolution of other genomes, or studies
of ribosomal RNA evolution can provide information
about the viruses.  HIV-1 evolves at a rate of approximately
1% per year within the ENV gene overall.  The Pol gene
evolves more slowly, most likley due to selection against
changes, not due to less frequent mutation.  I do not
know how the stem-loop regions evolve.  

*  Brian Foley          *  btf at                        *
*  HIV Database, LANL   *      *
*  Los Alamos, NM 87545 *  *

More information about the Mol-evol mailing list