Bioinformatics with Python Cookbook - Second Edition

There's more…

Is a reduction in error from 5% to 0.3% while losing 12% of markers good or bad? Well, it depends on what analysis you want to do next. Maybe your method is resilient to loss of markers but not to many errors, in which case this might help. But if it is the other way around, maybe you prefer to have the complete dataset even if it has more errors. If you apply different methods, maybe you will use different datasets from method to method. In the specific case of this Anopheles dataset, there is so much data that reducing the size will probably be fine for almost anything. But if you have fewer markers, you will have to assess your needs in terms of markers and quality.