Goal 2: On consensus and feature maps, the alignment approaches should find and group corresponding elements in the maps.
The TOPP tool MapAlignment solves both alignment problems. Our approach is based on the combination of pairwise map alignments. A pairwise map alignment proceeds in several steps. In the first step, the retention time warp and the distortion in m/z is estimated using a pose clustering approach. This initial transformation is used to find elements in the two maps which likely belong together. In the second step, these pairs are used as landmarks and a final, improved transformation is estimated by which the two maps are mapped onto each other in a third step. The first three steps constitute the so called superposition phase. In case of a feature map alignment, the corresponding elements are grouped together in a fourth step, the so called consensus phase.
In addition to the MapAlignment, we offer three tools for the superposition phase of a pairwise feature map alignment which are UnlabeledMatcher, MapMatcher and MapDewarper. These tools can be used if you want a fine-grained control over the matching process or if you are not interested in the feature pairs, but in the actual mapping function.
map_type
- The type of the input maps ('peak_map','feature_map', or 'consensus_map')matching_algorithm:number_buckets
- The number of buckets in retention time and m/z dimension. If the number is set to one, a globally defined warp is estimated and if the number is greater than one, the MapAlignment results in a piecewise defined transformation.matching_algorithm:superimposer:tuple_search:mz_bucket_size
- The deviation of corresponding elements in m/z.matching_algorithm:pairfinder:precision
- The deviation of corresponding elements in retention time and m/z after dewarping. Generated Tue Apr 1 15:36:40 2008 -- using doxygen 1.5.4 | OpenMS / TOPP 1.1 |