So what is common to transposon inactivation and paramutation is the following:
1. Silencing is associated with sequence duplication. In the case of transposons, this generally occurs as a consequence of transposition.
2. Promoter sequences become methylated preferentially.
3. Both transposons and R genes are transcriptionally silenced
4. Silencing is heritable, but reversible. When silenced TEs or R genes are reactivated, methylation decreases or disappears.
5. There is evidence of communication between homologous sequences located at different chromosomal sites. In the case of paramutation, the silenced locus can impose silencing on allelic loci, even when they are located elsewhere in the genome. In the case of TEs, a transcriptionally active transposon can activate an inactive one, leading to its heritable reactivation and demethylation. As I mentioned earlier, this is mediated by an element-encoded protein in the case of Spm. Another maize transposon, Mu, undergoes concerted methylation and inactivation of copies throughout the genome.
6. Finally, both involve transposon sequences. In the case of the paramutable R locus, a transposon sequence is integral to the promoter. The difficulty in interpreting this observation is that the deletion simultaneously removes the promoter sequences and the transposon sequences, so it cannot be ascertained whether paramutability and the ability to become paramutagenic requires the transposon sequences or transcription of these genes -- or both.