The difficulty of identifying genes in anonymous vertebrate sequences

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Gene identification in vertebrate genomic sequences presents significant challenges, ranging from straightforward to intractable.
• Current gene identification methods often rely on statistically "insignificant" individual features such as promoter signals, splice sites, and open reading frames.

Purpose of the Study:

• To evaluate the limitations of current gene identification methods in vertebrate genomics.
• To highlight the necessity for advancements in in silico gene analysis for effective genome interpretation.

Main Methods:

• Statistical analysis of individual gene identification features.
• Assessment of current popular gene identification methods, including database similarity search and GRAIL.
• Discussion of the feasibility of experimental characterization of human genes.

Main Results:

• Individual sequence features used for gene identification have limited statistical significance.
• Existing gene identification methods are conservative and hinder unexpected biological discoveries from genomic data.
• Automated interpretation of genomic data without experimental validation is currently not feasible for most cases.

Conclusions:

• The reliance on existing sequence databases and training sets limits the discovery of novel genes.
• Characterizing the vast number of predicted human genes experimentally is impractical.
• Significant advancements in in silico gene identification and analysis are crucial for transforming genome data into actionable information.