The mapping is automatically done at three level of granularity, i.e., at artifact template filter graph, at artifact template filter and internally to the artifact template filter at several variability points, such as attributes and methods when in the sync mode is selected or semi-automatically otherwise.
The experimental results show that the proposed approach outperforms the median filtering (MF) and the adaptive template filtering (ATF) in enhancing the MRI brain images.
Mapping at template filter graph level will be triggered by a meta-expression when at list two filters are linked in a stream.
Selected features of a filter will be successive collected till the moment that the filter can be evaluated by a meta-expression to a particular artifact template filter and the mapping realized.
Any other selected features will be mapped to internal variability points or attributes of the previously mapped artifact template filter by meta-expressions or presence conditions, respectively, indicating whether they should be patched or removed from the specialized template instance.
To take this into account, Weiner et al. identified peaks by template filtering using seven templates [9].
Current popular nucleosome calling methods are GeneTrack [ 126], template filtering [ 58], DANPOS [ 109], and iNPS [ 127].
However, the current version of template filtering is only suitable for small genomes (approximately 12 MB) due to memory limitations.
This file was then used in a Template Filtering algorithm with seven nucleosome templates [developed in Weiner et al. (2010)] to call nucleosome positions genome wide.
The template filtering algorithm was developed to control for the variable MNase cut patterns observed at different concentrations of MNase digestion.
Considering the default score thresholds, NucHunter and NPS return a similar number of predictions but the former has an higher AUC than the second, whereas Template Filter returns many more predictions and of lower quality.
