pileup question/bug?

[ml4wc]

Here is the entry in the vcf I get for the site in question:

chr10 11207607 . C G 255 PASS NS=1;CX=CG;N5=CACGC GT:GL1:GQ:DP:SP:AC:AF1 0/0:-6,-1000,-1000:255:10872:C5230A3G4T1Y5164:5234:0.00

The alternate is G and has a beta of 0.00, which makes sense the alignment visualization in tview and igv show a convincing mix of Cs and Ts, which is explained by the 5164Ys. In technical replicates for this sample and biscuit is calling this site correctly in those and reporting the beta. I also tried with and without -r option and nothing changed try for this dataset, still with a bunch of Ys. This might be my lack of understanding, but why is there ambiguity here and Y rather than Ts, which is what I see in tview and IGV.