Contributed by Jorge Dubcovsky

1. Spend some time identifying the conditions were the differences between for HIFs show maximum difference. For spikelet number photoperiod and temperature seem essential. We have found with Saarah that the differences are very compressed when we grow things in the GH in the summer… If you find a GH or chamber condition where you can differentiate your HIFs you will accelerate your project several years! If you depend on the field, do multiple environments, and try to identify the ones that are more reproducible. Always include several homozygous A and homozygous B as controls to be sure that the phenotype is visible in that environment.
2. Do enough replications in the progeny tests to Mendelize the phenotype. For the GPC-B1 we used 20 reps in the field, for the Rht18 we are using 100 plants in the field.
3. Adjust the number of segregating lines you study for recombination to get a manageable number of recombinants in each stage. If your interval is 5 cM you will have 10 recombinants per 100 F2. If you can handle 50 recombinants go for 500 F2. The number you can handle depends on the size of the Progeny Tests you need to run to get a clear Mendelian phenotype. Do the mapping in Phases. After a good mapping of this first 50 recombinant you can probably reduce the interval to 0.5 cM, so now to get 50 recombinants again you need to screen 5000 F2…
4. It is more important to have a few crossovers you are 100% sure, than a lot of recombination events that you are not certain. It is better to move at slower pace but with certainty. If you make a mistake in 1 crossover, you may end up walking in the wrong direction for a couple of years ( I did this twice!).
5. Positional cloning is not about measuring a large number of plants and running a QTL program. It is about cutting the region with crossovers and being 100% certain of the location of the phenotype relative to each of these recombination events. Do progeny tests of the size that need to be done to reduce the environmental variation to a point that you can determine the phenotype as a clear A, H or B allele. For each progeny test run an ANOVA using the genotypic data, if you have a significant segregation then it is an H. If no segregation is A or B. If you P values are between 0.4 and 0.12 repeat the experiment! You need to be certain.
6. The secret of positional cloning is to do your genetics well (PERFECTLY WELL I would say)
7. To reduce the genotypic variation, the more advance the HIF the cleaner the data. If you start with an F5 looking for recombinants, next generation  use the F6 HH to generate new and more advance HIFs, repeat each generation to keep purifying the genetic background.
8. If you are generating HIFs from an early generation, test 3-5 HIF families to see if there is a specific background where the differences are more evident. Then move ahead using that family. THERE ARE MULTIPLE EPISTATIC INTERACTIONS WE DO NOT KNOW. Your phenotype may be visible in one background but not in another one. You need to figure that out early on the game.
9. Genetic variability is  reduced  by sellfing or backcrossing, environmental variability is reduced by replication number in the PT.
10. Use the resources  we are providing. By now Eduard Akhunov has sequenced all the parental lines by EXOME CAPTURE. Jean Luc is working on tools to mine this data but you can get the raw data and do your analyses. Developing markers in the region is no longer a limitation!!!!