HiC Read Count Quantitation

The HiC Read Count quantitation is a simple way to look at the HiC coverage in a set of probes. It works very much like the normal read count quantitation, except that it counts at the level of HiC read pairs. Any HiC read pair with one end overlapping the current probe will be counted in the quantitation, but pairs where both ends fall into the same probe will only be counted once.

Options

The options you have for this module are:

1. Whether you want to correct for the total read count. If you are comparing different datasets with different total counts then this will normalise the average count per probe.
2. Whether you want to correct for the length of your probes. If your probes are of different sizes you would expect to get a higher count in longer probes. This correction adjusts the count to a count per kilobase of probe, so a 1kb probe would get raw counts and other sizes would be adjusted. Unless your average probe length is much longer than your average read length this correction could introduce a bias into your results since the correction is just off probe length, but the counts include any read which overlaps your probe - even if only by a single base.
3. If you want to log transform your count. If you have a large range of values in your count you can calculate them on a log2 scale to make them easier to view.
4. If you want to ignore duplicated reads. If you select this option then every unique read position (start, end and strand must be the same) will only be counted once and duplicates will be ignored

Warning

If you choose to quantitate your data on a log scale then any probes which contain no reads will have their initial read count increased to 0.9 reads to avoid problems with infinite values when log transforming. This value is set before any subsequent correction for total read count or probe size. This means that in log transformed data different data stores could end up with different absolute values for probes containing no reads. If probe length correction is applied then probes with no reads will end up with a range of low values reflecting the different lengths of the probes. In these cases you might want to use an initial linear quantitation to allow you to flag and possibly filter out the regions with no reads, before later moving on to quantitate on a log scale.