Hi, I received a variety of responses to my post, and I really thank all that responded. Copied below are my original post followed by the responses. If you are interested in this topic, there are some great suggestions here! To give you a bit more detail on my conclusions: -it seems the pattern is relatively common in at least some marine invertebrate groups. -I encountered this pattern in the last two pycnogonid species I have developed microsats for, using the Hamilton et al. protocol. With enough sequencing effort (~400 clones), I have successfully developed polymorphic microsats for them. So my query was intended to learn ways to minimize these problems in the future. -So in developing loci for future study species, I will be interested in incorporating some of the suggestions below, particularly with regard to decreasing the amount initial gDNA, using less-frequent cutters in the restriction step, and making sure to use probes tagged at the 3' end. Again, thanks a lot for the help! Felipe Barreto ORIGINAL QUESTION: Dear Evoldir members, I am currently isolating microsatellites for a marine arthropod, and have encountered a curious pattern. From a total of 300 clones sequenced so far, over 95% contain 'beautiful', long microsat regions, but in more than half of them, there is virtually zero (or very near zero) flanking sequence (i.e. the linkers/adapters used in the isolation protocol are followed immediately by the repeat region). I have anecdotally heard of this pattern in marine invertebrates, but have not found any clear literature on this. Has anyone experienced this issue to the same extent? Is this pattern real? Or am I possibly doing something wrong at some step (e.g. too much PCR product going into the sequencing reactions)? Thanks for any input!! Felipe Barreto Dept. of Ecology and Evolution University of California Irvine, CA ----- RESPONSES: Hi Felipe We made a microsatellite library a few years ago that had the same problem, and we realized it was because the initial restriction cutter we used had a recognition site that included GT, which happened to be the repeat we were trying to isolate. Maybe you've done the same? Cheers, Andrew ----- Dear Felipe, This is interesting question. I've recently encountered a similar problem with plant DNA during development of the microsats. We didn't really have 'zero' flanking sequences, nevertheless still so short that designing the primers was impossible or only possible with including a part of the microsat. therefore, I would be interested in the responses you get. Can you please summarize them and forward to me? good luck, Monika ---- Hi Felipe, You will probably get a load of answers about this, but I can tell you what we found in our lab. We had the same problem where we had beautiful repeats but the linkers were jammed right up against them. We thought it could be because there was too much DNA going into the initial enrichment procedure with the probe. The reason for this is that if there is too much DNA, the fragments that were lacking flanking sequence (just by chance) would be preferentially bound to the probe because the repeat region would be more exposed (i.e. there is no flanking region to "get in the way"). We quantified our DNA right before the enrichment procedure and started putting much less into the enrichment (unfortunately, I can't remember the exact amount right now and I don't have my notes with me, but I think it was around 100ng). This seemed to do the trick. We would still get the occasional sequence with no flanking region, but overall most of them were good. We had this problem across a wide variety of species (frogs, carnivores, birds, shrimp) and found that putting in less DNA worked for all of them. Another factor that could be contributing to it is the type of repeat you are enriching for and the restriction enzyme. We were enriching for GATA repeats, and using Sau3AI to initially digest the DNA, which has a cut site of GATC. So if one repeat had experienced a substitution from A to C, you could end up with a cut site right in the middle of your repeat. This might increase the number of fragments with no flanking sequence. We switched to HaeIII to digest the DNA (cut site GGCC). I don't know what your total protocol is, the only problem with this is that it's not recommended for use with SNX linkers (which we were using). I think there is a HaeIII cut site within the linkers, so if you have any residual enzyme it could cut off the linkers. To get around this, we did a single phenol-chloroform step after digestion to get rid of any residual enzyme. When we finally got good flanking sequences, we did see that sometimes there would be a GATC site within a GATA repeat, so I think it was definitely contributing to the problem. >From what I remember (this was several years ago!), the first thing we tried was switching the enzyme from Sau to Hae. I think it may have improved our results slightly, but not enough. We often got clones with one good flanking region but not the other. We then tried reducing the amount of DNA, and that seemed to fix it. It was very frustrating for us as well, and there does not seem to be any literature addressing common problems that are encountered when cloning. Hopefully this will help you! Kaela ---- I'm sure I won't be the first to mention this, and I can only say that it was our problem at one point. Anywho, be sure that whatever probe you are using to tag and grab the microsat fragments in your library are labeled on the 3' end, not the 5'. With 5' labeled probes, we discovered that in the PCR step of our enrichment process, where the enriched library is PCRed for ligation into the cloning vector, the excess probes began to act very effectively as PCR primers. So most of the cloned fragments appeared as you have described, with lots of beautfiul micro but frequently only one or no flanking regions! Stuart Willis ---- Dear Felipe, That's more or less normal in arthropods, hence it is very difficult to isolate microsats from a gDNA bank. They are simply too numerous to identify suitable flanking regions. A nice alternative is to make use of a cDNA bank, which gives you at the same time microsats and if you make use of several animals SNPs in coding regions. It's a bit more expensive, but in the long term really worthwhile. I attach one of our papers which was the first to isolate massively microsats in Penaeus shrimps. Good luck. Filip ----- Hi Felipe, Many years ago when I was a postdoc in Rick Grosberg's lab trying to isolate microsatellites from Hydractinia, I encountered the same issue and am sorry to report that we never got to the bottom of it. Brenda Cameron from Grosberg's lab and Rob Toonen (ex-Grosberg graduate student), may be able to comment more on this. They are copied on this email. Best, Alex Wilson ----- Hi Felipe, I've worked with marine crustacea msats libraries and found some long msats in some species of Decapoda. Which group are you studying? Your case seems to be that PCR fragments are too short (200-300bp maybe?) and you're missing flanking regions because of that. You should digest genomic DNA with different enzymes or shorter times. Did you use FIASCO enrichment protocol? Which repeat pattern did you used for enrichment and identifying colonies with msat? I wouldn't recommend dinucleotides, that's for sure. Cheers, Ferran felipesbarreto@gmail.com