How does RSEM allocate ChIPseq signals to genes?

Question

[moxu]

For RNAseqs, allocating the read seqs to genes is straightforward and easy: within the gene region or not.

How about ChIPseqs? For instance, if you are mapping Pol2 signals to genes, and you care about +-5k TSS regions, can you tell RSEM to do such an association? Even more complicated, if you want to bin the +-5k region into partitions, how can do do so?

Thanks.