Doubts in simple mapping technique

Hi

I had taken the cDNA sequence from a protein of interest and wish to find its promoter. The protocol I had adopted is as below:

1. I added the promoter features to the reference sequence using a GFF reader connecting to a repository

2. The cDNA sequence was then converted to a read format using a perl script and the depth was set to 40 by default

3. This read was then mapped to the reference sequence using BWA and the mapped reads were added to the repository

4. The mapped file [sam file] from the repository contents were then viewed with a tablet viewer with the output options set to reads in the repository query.

Results:

1. The coordinates of the promoters [start and end coordinate] in the GFF file are overlapping with each other and hence, the whole genome appears to be full of promoter sequences. How do i resolve this problem

2. The number of reads used as an input was one, but the number of output reads are 14 in number. How can this be possible. While mapping is there a cutting proceeduring adopted

3.  I collected the 14 reads and aligned them with the input cDNA sequence, but no regions in it are 100% similar. How can this be possible?

pls explain

I am sending herewith all the sequences:

file1.txt--mapped reads as observed from tablet viewer

file2.txt--input read sequence

I have used human genome as a reference genome sequence.