Why do annotating the same functionality with different databases produce different results？

Question

[Xudong Li]

Hello,

I made functional annotation associated with nitrogen metabolism for metagenomic data using KofamScan(https://www.genome.jp/tools/kofamkoala/) ，NCyc（https://github.com/qichao1984/NCyc）and eggnog-mapper. But these results are significantly different. For example, the ORFs that are annotatated as gene nirK using three methods are different. Why is it so?

Thank you.

Answer 1

[Jean-Karim Heriche]

Different resources may have different notions of what a gene is (they may use different references), they use different sources of information and collect and aggregate the information in different ways (for example by having different requirements/confidence levels). The result is that different databases contain different annotations.