Are introns just junk?

Eric E. Snyder eesnyder at beagle.Colorado.EDU
Mon Oct 25 16:47:40 EST 1993

cjp at (Christopher Pook (Bioc)) writes:

>apland at (Shannon) writes:

>>An article recently printed in the local paper stated: _2 win Nobel for 
>>finding most DNA is "junk"_.  The "junk"  they are referring to is the 
>>introns found in DNA.  

Not to mention intergenic DNA, which probably is a larger fraction of the 
total genome than introns (at least in higher eukaryotes). 

>An interesting idea that I heard somewhere, although I unfortunately can't 
>remember where, was that introns contain an error-checking mechanism which
>allows verification of the exon sequences flanking it. The idea came from
>a cryptologist/mathematician who tried "cracking the code" of introns and
>found a relationship between exon and intron sequences that allowed him to
>predict c.70% of the intron sequence from that of the exon. 

It is pretty easy to distinguish exon from intron given a reasonably large 
window of sequence.  Using standard backprop-type neural networks, Lapedes
et al. (JMB 226:471-479) can accurately classify 180 bp windows with a 
accuracy of "greater than 99%".  Of course that is a lot like saying that 
by traveling 100 miles of road I can distinguish between the Colorado Rockies
and Kansas.  The a more interesting question is can you tell the difference
between introns and intergenic DNA, which is more like saying can you tell the 
difference between Kansas and Nebraska.  

>He speculated
>that the apparently wasted energy that the cell puts into transcribing
>precursor RNA (if the introns role is in DNA recombination or replication
>then why include them in RNA??) allowed the cell to verify the transcribed
>sequence before translation, thus preventing synthesis of incorrect protein.

Verify the sequence before translation?   There was a paper in Science a 
few months back that showed a case where a single nonsense mutation in an
exon altered the pattern of mRNA splicing.  Very interesting... they 
proposed no mechanism for this effect but it certainly makes one wonder.

>This system would therefore be very similar to error-checking protocols used
>in communication between computers, where extra "junk??" bytes are added to
>the data, their value being determined by some property or relationship with
>that data (eg. if the sum of each bit is odd/even add a 0 or 1 respectively).
>This allows the recipient to check for mistakes and, I believe, more sophis-
>ticated methods can even correct any mistakes detected.

My last comment aside, I think there is little evidence that introns are 
involved in any sort of proofreading or error correction activity.  Instead,
I think introns were introduced via transposons or retroviruses and eukaryotic
cells managed to find a way to cope with the new headache.  Eventually, 
metazoans learned to exploit sloppy splicing for developmental and other
types of regulation and managed to get really good at it when a single 
specific product was required.  

Just to bring it up again, has anyone really studied the difference between
intergenic and intronic DNA.  I know of one paper by Konopka that looked at
the local complexity and periodicities of 5'- and 3'- regions of genes but
that is all I can think of...

Eric E. Snyder                            
Department of MCD Biology              ...making feet for childrens' shoes.
University of Colorado, Boulder   
Boulder, Colorado 80309-0347

More information about the Cellbiol mailing list