How to get the exact mismatch positions in Tophat output

[maplesword]

I ran TopHat 1.0.14 and got the mapping results. Now I'm trying to find out which positions are mismatches when mapping to the reference genome for each read. The sam output of TopHat is like this:

[ID]  16  chr1  14509  0  76M  *  0  0  [Query_Seq]  [Quality_Score]  NM:i:2  NH:i:7  CC:Z:chr12  CP:i:91041

I checked the format manual of SAM. "NM" in TAG part shows the number of differences between the query sequence and the reference sequence. That means in my example there are two mismatches. However, I cannot find out which two positions of this read are the exact mismatch ones in this output. In SAM format, there is one kind of TAG "MD" which shows the mismatch positions but I cannot see any in the output of TopHat. Can anyone tell me how to do?

Thank you very much!

[maplesword]

I think I have at least solved parts of the problem with samtools.
First I used "import" command of samtools to transform the sam file to bam format, and then used "fillmd" command to add the MD tag.

But it comes another problem. When adding the MD tag, the program also output lots of warnings showing that the NM tags of lots of reads are wrong. I don't know why. Can anyone give me help? Thank you!