Reconciling TN sample list and our transcribing

Alex Buerkle used the complicated TN spreadsheet to create a list of tubes that need to be amplicon sequenced. Shannon Harris had already arranged the tubes they had sent and transcribed the hand written labels into a spreadsheet and added them to 96 tube racks. I attempted to reconcile the tobeplated list from Alex’s data manipulation with this using this r-script and these data. I was able to locate 293 of the 303 tobeplated tubes and created a pull list. I found one transcription error. There are still 10 tubes I could not locate:

I tried to delve into the data to see if I could locate these 10 samples. I did not have any luck. I do not think it is a transcription error when I anti joined samplesInSites against our spreadsheet and then the spreadsheet against samplesInSites there was no easily discernible pattern. The list of sites between the 2 even disagreed. And, there were 212-228 disagreements.

212 samples from the transcribed list do not match anything in samplesInSites (an R object that reflects the spreadsheets that came from Tennessee). These are in DisputedTubes in the R workspace.

To make progress understanding where these samples might have come from, I looked at what there sample type was:

table(gsub("^Site_[[:digit:]]+(.*)[[:digit:]]+$", "\1", DisputedTubes$Sample))

       _Alf_Epi             Alf_Epi Site_129_Alf_Epi_5_                Soil 
              1                  89                   1                  16 
     Trans_Endo           Trans_Epi 
             74                  31 

Most of the samples are from sites 105, 106, and 107.

table(gsub("^Site_([[:digit:]]+)_.*$", "\1", DisputedTubes$Sample))

101 103 104 105 106 107 108 109 110 116 120 125 126 129 134
1 10 10 40 39 40 4 15 14 10 2 7 9 2 9

So, something is still fishy with the TN spreadsheet concerning tubes.