Automatic multigenic family annotation: risks and solutions

Abstract

A major challenge facing bioinformatics today is the efficient annotation of the exponential flow of genomic data. This has led to an increasing dependence on automatic annotation procedures, despite the relatively high error rates of these programs, particularly for multigenic families. We discuss here the errors and biases introduced by automatic genome annotations, focusing on issues with structural annotations of gene families, and suggest ways to overcome these limitations.