Webinar

AlphaFold for gene structure annotation

Webinar overview. The Rigden group uses AF and other protein structural bioinformatics tools to address diverse challenges, specialising in the translation of AF for novel purposes in neighbouring areas of biology. In joint work, co-led by Prof Andy Jones, we explored whether AF could help select the correct gene structure model, using original and manually curated pairs from two fungi and an apicoplast. We found that where two alternative gene models led to AF predictions that differed in a structured domain, structure-derived scores strongly favoured the human-curated version. We further found that the structure-based signals were complementary to sequence domain analysis. Unsurprisingly, the AF approach was not helpful for gene structure uncertainty relating to potentially disordered protein regions, highlighting the need for further research. We explored faster tools as alternatives to AF3 and found that Protenix-mini retained the same discriminatory capacity. Since publication we have further explored scoring of complete proteomes and have built a NextFlow pipeline to facilitate adoption of the approach.

 

Who is this course for?

Anyone interested in unexpected applications of AlphaFold.

Anyone looking to predict gene structure better in eukaryotes, especially non-model species.

Outcomes

By the end of the webinar, you will be able to:

  • Understand the ability of AF to score candidate gene models
  • Perceive the complementarity of the AF-derived scores to other metrics
  • Understand the value of AF-derived scores where transcriptomics data are lacking
  • See how a NextFlow pipeline should help you apply the approach 

EBI Resources

title
19 November 2026
10:00 – 11:00 ( GMT )
Online
Free
First come, first served
1000 places
Contact
Sheila Obilo

Organisers

Speakers

Share this event with: