Thank you to everyone who replied!  I was able to reduce the ambiguity
in identifiying the 'true' alleles by implementing suggestions for PCR
optimization. In particular, reducing the number of cycles and the
time at each step, resulted in PCR products in which the stutter was
reduced.  Previously, each allele at a heterozygous individual, had
exhibited several peaks of very similar size.  After reducing the
number of cycles and the time at each step, each allele had one larger
peak with several smaller stutter peaks in front of it.

There were numerous requests to post the replies so here they are:

Here is a repeat of a post Brian made some years ago compiling responses to
a similar question.  Hope it helps.
RESPONSES TO MICROSAT STUTTERS
Below I have grouped the responses so far received to reduce stutters,
thank you to everyone who has responded! The most common response was to
reduce the number of cycles. Followed by reduce the MgCl concentration.
Also a mix of PCR cycle options with a high initial annealing temperature
either through a 'Touch down' process or a few said plain heating your PCR
before you put the polymerase in. 'Two-step' PCR was mentioned more than
just heating your reaction without taq. Also as popular was to reduce the
extension time at the end. About 3-4 people said specifically to try hot
start taq. After the combined PCR suggestions commonly stated was to just
redesign reverse primers according to the 'Pig-tail' method (Brownstein et
al. 1996, Biotechniques 20, 1004-1010).
Since then I have reduced the number of cycles, then also shortened the
extension time. Then with the shortened number of cycles and extension
time varied MgCl from 1.5 to 2.5 because the person who uses them
successfully now says not to go out of this range. I have also kept the
2.5 MgCl as original, but added BSA. I am currently looking over the
results, but wanted to pass along the feedback I have been getting.
Again THANK YOU very much to all who responded.
PCR REACTION CHANGES
. Varying Mg++ levels, maybe try a gradient from 1-4mM, lower MgCl helps
to reduce stutters.
. Use hot start Taq
. Change concentration of your primers. We use a starting concentration
of only 0.5uM of our labeled primer and then the reverse (unlabeled) is
2uM. My labmates have found that better bands sometimes come with changing
the labeled primer to either 0.3uM or 0.7uM. We were told that it's good
to have the reverse primer in higher concentration.
. Try out t4 polymerase
PCR CYCLE CHANGES
. Reduce the number of cycles down as low as possible, 40 cycles is
excessive and try 25 cycles
o "increasing the number of cycles may increase nonspecific background and
may lead to artifacts" (source: Qiagen multiplex pcr handbook, 2002, p.33)
. Use a very stringent annealing temp
. Set your final extension to 70 degrees or 65 degrees for 45 mins (see
Profiler Plus manual for Profiler Plus kit used on humans). I had some
loci which were tricky to score until our lab stumbled across the reaction
conditions used for Profiler Plus kit.
. 'Touch down' PCR program
o start Tm at 65°C or 67°C and the annealing temp reduced 2°C per cycle
for the next cycles up to Tm=58, then the remaining cycles at a 58°C
annealing temperature
o start at 60°C and within the first 10 cycles down the Ta to 50°C, then
keep 50°C for the next 30 cycles
o start at 60°C and within the first 10 cycles down the Ta to 50°C, then
keep 50°C for the next 30 cycles
. "short PCR"
o cycles of 15s at 94, 15s at annealing temperature and 45s at 72 and
lower the number of cycles. I've heard people sometimes run only 5s at
annealing temperature
. Hotstart PCR with regular promega taq by performing a first denaturing
step with DNA, primers and water and then lower the temperature for about
85ºC to add the master mix containing taq and all the other reagents. I've
performed 12.5 ul reactions, primer concentrations ranging from as low as
0.03 to 1.2uM in multiplexes. Annealing temperatures were 58ºC for most
primer sets and 60ºC. I also recommend the Quiagen multiplex PCR mix,
that provides a taq appropriate for hotstart PCR.
. No 72°C extension step in the cycle (I mean that I have the annealing
step followed by the denaturing step and then annealing again, with the
Taq extending during the ramp, for microsats is more then necessary), and
only 2 minutes of final extension step.
. Biotechniques suggested "abbreviated cycling times" for PCR steps. For
instance, denature for only 2 seconds, anneal for 2 seconds, extend for 8
seconds. I've often used this "2/2/8" protocol (or some variant, e.g.
3/3/12) to clean up microsatellite stutter. Sometimes it fails miserably,
but in about 50% of the cases it helps dramatically.
. Try heating your PCR without taq in it for 15min @95C,then adding taq
. Two-step PCR
o 30sec at 94C and 80 sec at 50C. Also here is the reference that uses
the suggested cycle Here is the reference:
 YoshiadK, Awaji M (2000). Improvement of polymerase chain
reaction condition to detect polymorphic dinucleotide repeat
microsatellite DNA marker in the puffer fish. Fisheries Sciences, 66(2),
397-399.
o denaturing at 94 and annealing at 62-65, with no elongation step- this
sometime helps
PRIMERS
. Re-designing the reverse primers the way that you add GTTT to 5'end,
i.e. if your original primer has A,G,C, at 5'end add GTTT, if it ends by
T, add GTT, if by TT add just GT etc. You may find info in Brownstein et
al. 1996, Biotechniques 20, 1004-1010. Also see ABI Prism Linkage Mapping
Set User Bulletin for 5 Nov 1997 re: "plus A artifacts".
ADDITIVES
. There are several additives you could try adding. They include dimethyl
sulphoxide (DMSO, up to 10%), formamide (up to 1%), tetramethyl ammonium
chloride (TMAC), BSA. The formamide and TMAC can be used together.
OTHERS
. Fresh formamide (formamide should be double-deionized, aliquoted as soon
as it's received, and frozen). Thaw and use small aliquots and throw
them away if not used up in about 2 weeks.
. Quality of the DNA, run your genomic DNA on agarose just to see that it
is not sheared, degraded or of low template amount. Also Decrease the
concentration of your DNA.
. Run your results through Micro-Checker (website address:
http://www.microchecker.hull.ac.uk ), and see whether the pattern is
consistent with deviations due to stuttering. (This assumes that you do
population genentic analysis on single populations that are in Hardy
Weinberg equilibrium. I could extract from your email what your research
is about, so perhaps that assumption is violated).
. The membrane combs can give different mobility than loading into wells
made by sharkstooth combs, try sharkstooth combs to see what you get.
. I was unsuccessful in ever finding a way to reduce the stutter peaks.
. Run your loci on a very fine and dense agarose gel (or polyacrilamide),
cutting out the band you think to be the microsat and sequencing it. Then
you will know for certain the microsat is there
. Maybe your target species is a recent polyploid or hybrid and you are
picking up multiple copies of each msat locus
. Stutter is unavoidable with most microsats. The usual pattern is a
series of peaks, each one separate from the others by a number of base
pairs equal to one repeat unit, with the peaks increasing in height/area
as they increase in fragment length, up to the highest peak. Then there
is a sharp falloff, with one or two peaks much lower than the highest one.
The highest peak represents the actual allele. On the other hand, if
you're getting a number of adjacent peaks of similar strength there's
probably something wrong. This is particularly true if your products
aren't amplifying strongly. It shouldn't take more than 0.5uL of PCR
product to give you a decent-sized peak on an automatic sequencer. If
you're getting a weak PCR and the primers aren't working.

You might try greatly abbreviated times at each step in PCR (e.g., 2 sec
denaturation, 2 sec annealing, 8 sec extension).  The original paper
describing this was published in Biotechniques about 6 or 7 years ago;
see also the attached reprint.  [We had cited the Biotechniques paper in
the bass reprint, but the citation was ultimately cut due to strict page
limits in PRS.]  In my experience, the approach either works very well
or fails completely (on a locus by locus basis).  If you get
amplifications, the stutter is usually greatly reduced.
DeWoody et al. (2000) Proc. R. Soc. Lond. B 267:2431-2437.

you may find helpful reference in the attached review: Pompanon et al.
(2005) Nature Reviews Genetics 6:847-859.

I have once heard that stutter is produced in PCR and that some
DMSO in the PCR reaction might reduce it.
Another possibility is that denaturing gels (as are usually used) are
producing this stutter. I have once worked with pre-cast high-
resolution gels, I think they were from GATC Biotech, a German
company (http://www.gatc.de/en/index_en.php), but there might also
be other suppliers. However, these gels are horribly expensive and
not very convenient to run (temperated elpho chamber, staining with
SYBR Green necessary etc.). But on these, I saw only a single band
in a homozygous individual whereas the same sample produced a
stutter on denaturing polyacrylamide gels.
Instead of trying to reduce the stutter, you may define a way how you
interpret them and maybe run reference samples as internal
standards in cases of doubt side by side with the samples of interest.

You could try PIG-tailing (adding the sequence GTTTCTT to the 5' end
of one of the primers), to ensure an A is added to the end of all PCR
products, thereby reducing stutter. See paper:
Brownstein, M J, Carpten, J D, Smith, J R (1996) Modulation of
Non-Templated Nucleotide Addition by Taq DNA Polymerase: Primer
Modifications that Facilitate Genotyping Biotechniques 20:1004-1010

I have yet to find PCR conditions that will help reduce stuttering
patterns.  We have not had luck changing the magnesium concentration in our
tests.  Stutter patterns may be reduced by primer design, but I have no
experience testing this.  I do find that using higher quality reagents
(like HotStar Taq or Qiagen's Multiplex PCR kit), which more costly
up-front, produce a more consistent pattern, which may in turn make
interpreting stutter-prone loci easier.  So, in my lab we recommend
re-amplifying and re-scoring samples displaying potentially problematic
peak patterns (e.g., indicative of adjacent-allele heterozygotes) in order
to estimate error rates.  Sufficient screening of stutter-prone loci and
comparison of patterns among reanalyzed samples may allow near-neighbor
heterozygotes to be distinguished from homozygous peak patterns.  In this
case, careful and consistent manual scoring of electrophoretic patterns may
reduce the number of scoring errors in the final data set.

stutter can be pretty annoying at times. I would assume you will always have
to deal with a little bit of it, but generally, stutter can occur because of
two reasons: a)Enzyme slippage in the PCR - and therefore loss of alleles in
the PCR. Solution: Add PCR additives to help the Taq over this difficult
regions. Additives can be such as NP-40 (=Triton X-114, 0.05% conc.) or
betaine. BSA helps to keep the Taq in better shape over time.
b) A-tailing of PCR products. Taq adds an A to the end of your product. If
this is not happening to completion, you will see lots of stutter. So, set a
good long final extension step, and play a bit with your PCR conditions.
In addition, longer repeats, such triples or tetras will make your signal
much clearer. Make also sure that both your primers anneal equally well.

The presence of double peaks or more looklike stutters,are most likely
due to incomplete adenylation of my fragments by Taq polymerase,
producing A overhangs and resulting in two peaks for each allele,
which then differ by a single base.  You can solve this by either
taking away the A from all the peaks or forcing everything to be
plus-A, which is the easiest.
The most prominent way to get over this is by Pigtailing your
unlabelled primer at the 5' end (Brownstein et.al 1996 BioTechniques;
Magnuson et.al 1996 BioTechniques). This should lead to complete
adenylation of all fragments.
 Other suggestion that could be tried out are
1. Add a 45 minute to 1 hour step at 60 degrees (alternatively 72oC)
immediately after the last cycle (So no final extension at 72
degrees). This step provides no free nucleotides left. All amplicons
will have an A overhang (plus-A).
2. Use less template DNA.
3. Tryout Touch down PCR
4. Try Different MgCl2 Concentrations ranging from 1mM to 6mM.
5.  After PCR completion, leave your PCR products at room temperature
O/N or Deep freeze it.
6. Use Mungbean nuclease to get –A molecules by polishing the product
7. Usage of Pfu polymerase enzyme or a combination of Pfu and the
normal Taq (As Pfus are expensive) to get rid of A overhangs during
amplification (thanks to the proofreading activity of Pfu)
8. Use Invitrogen´s Tsp DNA polymerase or GENOTYPE Tsp DNA polymerase
by Life technologies (which lacks terminal transferase activity)
9. Perhaps you can also dilute your PCR product before loading it to
the capillaries.
10. Add 0.4U T4 polymerase to 10ul of your PCR product, incubate at
37oC for 30 min, in order to obtain equally sized fragments.

katieshul@gmail.com