I have hundreds of humans who have undergone SNP genotyping at hundreds of loci. Some have even undergone the procedure twice or thrice (kind of an internal control).
So obviously I need to find those replications, and confirm that the results are the same. If there is discordance then I need to address it. I tried to use the aggregate function nr.attempts <-aggregate(RawSeq$GENOTYPE_ID,list(sample=RawSeq$SAMPLE_ID,assay=RawSeq$ASSAY_ID),length) This was simply to figure out how many times the same piece of information had been obtained. I ran out of patience. It took beyond forever and tapply did not perform much better. The reshape package did not help - it implied one was out of luck if the data was not numeric. All of my data is character or factor. Instead I used RODBC sqlSave(channel,RawSeq) to push the table into a Microsoft Access database Then a sql query, courtesy of the Microsoft Access Query Wizard a la design mode. SELECT RawSeq.SAMPLE_ID, RawSeq.ASSAY_ID, Min(RawSeq.GENOTYPE_ID) AS MinOfGENOTYPE_ID, Max(RawSeq.GENOTYPE_ID) AS MaxOfGENOTYPE_ID, Count( RawSeq.rownames) AS CountOfrownames FROM RawSeq WHERE (((RawSeq.GENOTYPE_ID)<>"")) GROUP BY RawSeq.SAMPLE_ID, RawSeq.ASSAY_ID ORDER BY Count(RawSeq.rownames) DESC; This way I could easily use the minimum and maximum values to see if they were discordant. Microsoft Access handled it with aplomb. I plan to use RODBC to bring the result of the SQL query back into R. This is the first time I have seen Microsoft Access outpace R. Is my observation correct or am I missing something. I would much rather perform all data manipulation and analyses in R. -- Farrel Buchinsky [[alternative HTML version deleted]] ______________________________________________ R-help@stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.