cycle vs. normal sequencing

Thu Dec 22 10:48:02 EST 1994

Clemens Suter_Crazzolara writes:

> I asked this question once before, however got only very few
> answers. How does, in your hands, cycle sequencing compare
> to the classical methods ?

I don't remember if I responded before, but here's my opinion:

On the plus side, cycle sequencing does much better on templates with
secondary structure problems, solves problems due to secondary priming
of the sequencing primer hence allowing larger templates, and can
address smaller amounts of DNA hence allowing direct sequencing from
colonies or plaques.  It's the only method that works well on small
linear ds templates, IMHO.

On the negative side, it is technically more cumbersome, slower,
expensive to scale up, and gives shorter reads due to more background
bands in all positions.   IMHO, internal labelling with S35 doesn't
work well enough to pursue, in contrast to the classical method.  So
at large scale, you end up with more P32 in use than you'd like.  P33
solves this problem, but it doesn't work as well on very small amounts
of template as P32.

So it depends on what you're doing.  Cycle seq. is clearly terrific
for direct screening of colonies or plaques by seq. because you can
leave out the template preps.  It's also great for direct seq. of PCR
fragments because you can avoid cloning them and the subsequent
template preps.  

On the other hand, if you want a few kb of contiguous sequence off a
single clone, then the classical method coupled to short medium and
long gel runs milks more sequence out of each primer.  And presumably
this DNA was important enough to you that you made an ample supply of
it anyway, so the logistics of template preps are irrelevant.  So if
you do a lot of sequencing in this mode, you probably want to gather
most of the data by the classic method, and fall back to cycle
sequencing for compressed regions and troublesome primers.

Now suppose you're into mega scale sequencing with a shotgun strategy
to start followed by directed sequencing to fill in holes.  Since
the required redundancy in shotgun strategies plummets with longer
reads, I'd prefer to do that part by classical methods.  When you get
to the hole-filling stage, cycle sequencing has the distinct advantage
that you may be able to prime directly off of large templates (like a
cosmid) thus avoiding a host of ad hoc cloning projects.  However, if
you plan to plow through repetitive sequences within the cosmid
insert, you probably have to subclone them anyway to get unique
priming sites.

My comments above are slanted towards sequencing with radiolabel.  If
you use automated seq. with fluorescent primers, then there are some
more issues.  The advantage of the classical method for longer gel
runs is blunted because the machine doesn't do well beyond 300 bp from
the primer anyway.  The key issue is that the machine is less tolerant
of poor quality ladders than is X-ray film.  So you have to be leary
about whether cycle seq. can routinely produce sufficiently clean data
for this method.  Some people clearly can do it, but I can't judge yet
whether routine coupling of cycle seq. to the machine is a good idea.

Steve Hardies, Dept. of Biochemistry, Univ. of Texas HSC at San
Hardies at

More information about the Methods mailing list