I imputed my data with TopMed imputation server, and now I have 3 problems
1- I have no more rs ID, all variants are chrXX:00000000 format. How can I resolve this, I am working on vcf and plink files
2- If I allply an r² threshold, could I keep the physically genotyped variants or it will keep only the best imputed as genotyped variants have no r² ??
3- The results from genotyped data showed a strong association signal, but this signal disappeared with the imputed data. Could it an imputation error ??
Did you do any pre-imputation QC, and specifically did you run something like Will Rayner toolbox to prepare data for imputation to HRC or 1000G panels? Check it out here: http://www.well.ox.ac.uk/~wrayner/tools/. TOPMed imputation server documentation seems to imply that you just need to run checkVCF, but ...
It could be that your genotyped variant was dropped because it did not "match up" with what was in the TOPMed panel. Could be SNP name or REF/ALT problems etc. Just curious because I am using Will Rayner's updated tools to check my data. I think his instructions are for Freeze 5 data though and the imputation server uses Freeze 8.