Home | Datasets | References

Nucleo Version 1.0 was trained to distinguish between proteins imported into the nucleus and those that are not. The dataset was curated from Swiss-Prot Release 50.0, and includes both permanent nuclear proteins and shuttling proteins with dual localisations. The training set consists of 2842 positives and 2606 negative samples added to Swiss-Prot up until the start of 2005. The training data are available upon request.