You'll need to do microassembly (e.g., with mapsembler) using the transgene sequence. The biggest problem with this will be if your ChIP signal is very sharp or not near a transgene edge then you won't get enough coverage to find the insertion site(s). Just google around for microassemblers. You will also have problems if you have too many insertion sites (e.g., you're sequencing a highly diverse population), but there's no way around these issues given the data you apparently have.