Closed greatfireball closed 6 years ago
Since nobody added a comment to that issue, I discussed that issue with @iimog directly. We decide to evaluate the tools using different simulated data sets. That should be done in parallel to the analysis of real word data sets. I created a new issue #6. Further comments there.
Do we need to test all candidate programs with different ratio for known chloroplast/genomic reads? Therefore, we could use a read simulator and species with known chloroplast and genome sequence. Different ratios should be used. This will test each program for its suitability to assembly whole chloroplasts from a read mixture.
We already did this analysis for an old version of chloroExtractor, but not using the current version.
Opinions?