Are tags the same as reads

Question

[gamore]

Hi, I made a Tag Directory (using HOMER's makeTagDirectory) from a BAM file. When I compared the tagInfo.txt to the statistics of the BAM file (first sorted the file using samtools sort and then samtools stats) the numbers don't match: in the BAM the 'reads mapped' are 3273271 and in the Tag Directory the 'total tags' are 1454419.5. I thought tags and reads were equivalent terms used for different formats, but in that case the total tags number shouldn't have decimals. I would appreciate some information on what tags are exactly and how they are determined (couldn't find it in the manual). Thanks!

[GenoMax]

Have you checked this page?

[John]

From genomax's link:

HOMER can use paired-end reads. If you use paired-end reads for ChIP-Seq or RNA-Seq, HOMER will treat each half of the read separately (and count each as 0.5 reads), which works well for a number of applications.

A number of applications... hehe - well i guess 0.5 is a number. I have no idea what HOMER is doing in the BAM to HOMER conversion step, but it's probably filtering reads out that don't pass certain requirements. Secondary alignments or low MAPQ values. FWIW, TagDictionaries is a really awful name for their data format, since a tag is a well defined thing in the BAM file and nothing to do with TagDictionaries at all.

Good luck.