Orphelia [Datasets]

Datasets for download

We provide the training data in Matlab® vector format (*.mat). The neural net is based on Netlab structs. For evaluation the test fragments are provided as plain sequence files (*.seq) with the corresponding predictions by our method (*.coords) and genomic coordinates of the fragment (*.cuts).

• detailed description of the data and the model (pdf)

• prediction model (Matlab-file, 59 k)

Training data

• discriminant (monocodon/dicodon) (Matlab-file, 255 MB)
• discriminant (tis) (Matlab-file, 155 MB)
• neural net (Matlab-file, 489 MB)

Test data

• test fragments (*.tgz, 1.2 GB, decompressed 5.6 GB)

Please direct your questions and comments to orphelia-info@gobics.de.