gene annotation in sequenced cancer genome

[SF_mallish]

Hi all,

I am dealing with ChIP-seq data for a small lung cancer sample. The genomic sequences of this cancer sample is available in this paper: "A small-cell lung cancer genome with complex signatures of tobacco exposure". We are trying to build a specific reference genome based on the whole genome sequences provided in this paper and map our ChIP-seq data back to this specific reference genome.
My problem is that the gene annotation files like GTF files will be very different for this specific reference genome and common reference genome (Hg19), because of the somatic variation (insertions, deletions, rearrangements) in this cancer genome. The coordinates for transcripts will change a lot. Is there a software of program to solve this problem?
Something like by providing hg19 GTF files and information for all somatic variations as listed in this link, it can output a gene annotation file for this specific cancer reference genome.
or if not, could someone give a clue about how to cleverly do that?

Thanks a lot!

[Richard Finney]

If this is human (I'm assuming it's not one of these fellows : http://www.holytaco.com/25-smoking-monkeys/ ), you can get the sequences for the human genes from genbak/ncbi/entrez and blat them against your custom(?) genome. This is fairly easy to set up if you know how to script and process a big data set. It is still a big job in terms of horsepower needed but easily paralleizeable if you have access to many computers. You may need to fine tune the blat alignment parameters and filter the results for good hits. The results will give you the coordinates for the genes.