partial methylation/intermediate methylation

[lancelothk]

For single CpG site, what may cause the partial methylation/intermediate methylation in BS-seq data?
What I know are:
1. allelic methylation difference in diploid cell.
2. sequencing/bisulfite error (after trim lower quality bps, filter out low bisulfite conversion rate reads)
3. Inter-cell variation for mix cells sample.

How to distinguish this from other type?
Is there any literature about this topic?

Thanks.