Aligning paired end reads against a BAC

[gen2prot]

Hello All,

This is a question about annotation improvement using paired end RNA-seq reads. The following is the dataset I am working with.

1) Sequenced BAC from Species A cultivar A: This is what I am annotating.
2) Gene model transcripts from Species A cultivar B: Dataset that is getting aligned to the sequenced BAC and I have high confidence about its quality. So this is my gold set.
3) Whole genome transcriptome paired-end reads from Species A cultivar A: This is what I want to use to improve the annotation of the BAC, for example making gene model edits. Also of high quality.

My questions are,

A) Should I align the paired-end reads directly to the BAC sequence using e.g. Bowtie, visualize in Apollo along with the other evidences such as (2) and then make corrections to the questionable models OR

B) Should I align the paired-end reads to (2) first, then take these gene models and align to the BAC using e.g Gmap

C) Is it advisable to take Whole genome transcriptome paired-end reads and align to a genome fragment such as a BAC.