Page 1 of 1

Number of training genes too low

Posted: Wed Sep 16, 2020 11:21 am
by AMorales
Good morning,
I am trying to run a training in Augustus, for that I upload the genome file of my organism, which does not have a pre-training parameter set. Since my organism belongs to a not very studied or popular taxa, there is not much biological information, however, there are some transcriptomes reported at NCBI that I assembled and used as cDNA to submit along with the genome file. When the training is submitted, the next error appears :

Number of training genes is with 34 too low (at least 100 genes required)! Training aborted.
failed to execute: perl /usr/share/augustus/scripts/autoAugTrain.pl --cpus=8 -t=/data/www/webaugustus/webdata/augtrain/trainCVxrC4Wj/autoAug/trainingSet/training/training.gb -s=trainCVxrC4Wj -g=/data/www/webaugustus/webdata/augtrain/trainCVxrC4Wj/autoAug/seq/genome_clean.fa -w=/data/www/webaugustus/webdata/augtrain/trainCVxrC4Wj/autoAug -v -v -v --opt=1

And I'm affraid that the reason why the number of training genes is too low depends directly on the cDNA file, is tha rightt? Since it could be a troublesome fact because there are not many other options of cDNA to try. What could be the main reasons behind this error, and possible solutions?

Thank you so much in advance!