Section: New Software and Platforms

magecal

Keyword: Genomics

Scientific Description: Magecal independently runs training and prediction steps for Augustus, Conrad, GeneID, GeneMark, and Snap. The results are cleaned and integrated into a common format. Jigsaw is trained and used for model reconciliation. Consistency constraints are applied to ensure that phase and intron structure are biologically plausible.

Functional Description: Magecal predicts a set of protein coding genes in fungal genomic sequences, using different de novo prediction algorithms, and reconciling the predictions with the aid of comparative data. Magecal applies consistency constraints to guarantee that the predicted genes are biologically valid.

Release Functional Description: Dockerization and compatibility with Alcyone

• Participants: Pascal Durrens and David Sherman

• Contact: David Sherman

• URL: https://gitlab.inria.fr/magecal/magecal