JEDI paper's modified version of ampliseq

In v0.0.1 and v0.0.2 a number of read threshold was used for each sample to decide if the consensus merging strategy was applied or not. The issue was that the threshold was applied to the number of rows in the merger table which correspond to a number of ASVs rather than a number of reads. So with a threshold of 1,000, a lot of samples were not merged with the consensus approach.
The motivation behind such a threshold was to have enough observations when defining the overlap size threshold, which is the first permilles of the overlap size distribution of ASVs retained by the default merging strategy. Now this value is defined at the run level and the number of read threshold is not relevant anymore.

v0.0.2

fix one sample/run bug

Consensus strategy prototype

v0.0.1

increase min number of reads consensus 1000