Help Documentation

What is a NoLS?

NoLSs are nucleolar localization sequences. NoLSs are short basic protein motifs that localize the proteins that harbour them to the nucleolus, a sub-compartment of the nucleus. A number of NoLSs have been experimentally identified and validated in human proteins. Many of these NoLSs can target non-nucleolar reporter proteins to the nucleolus.

Back to top

Sequence-based search

NoLS predictions can be obtained by pasting a protein sequence in fasta format in the input box and pressing the submit button. Optionally, users can decide to consider Jpred secondary structure prediction for the prediction of NoLSs. This results in more accurate predictions but requires more computation time. If possible, full-length protein sequences should be used to obtain maximum prediction accuracy.

Once the protein sequence has been submitted, a waiting page is displayed providing users with a link to the output page. This link can be bookmarked and consulted later. The waiting page automatically displays the output once the prediction is completed.

Explanation of the output

Back to top

Accession-based search

Our database of predicted NoLSs in human proteins can be searched by entering IPI, RefSeq or UniProt accession numbers.

Explanation of the output

Back to top

The predictor

NoD is an artificial neural network predictor built using the Stuttgart Neural Network Simulator.
NoD considers as input the protein sequence (windows of 13 residues) as well as protein length, position of the window within the protein and secondary structure predictions from Jpred when available.

As shown in Figure 1 and Figure 2, for each 13-residue window of the protein sequence, the input is encoded as a vector of 165 numerical values which are fed to the neural network, which in turn generates one output value, a score s between 0 and 1. This score is assigned to the middle residue of the window.

Figure 1

All 13-residue windows of the protein sequence are assigned an output score. NoLSs are predicted as regions of the protein in which at least 8 consecutive windows have an average score of at least 0.8.

Figure 2

Back to top

Explanation of the output

For proteins predicted to contain at least one NoLS, the output consists of 3 sections:

  • The top section enumerates the sequences of the predicted NoLSs
  • The middle section displays the input sequence with the predicted NoLSs shown in red
  • The bottom part displays a graph of the NoLS window-based score as a function of position in the protein sequence

    Figure 3 displays typical NoD output and predictions for the protein NOL12:

Figure 3

The score shown in the NoLS score per residue graph represents the average score of a segment of 20 residues representing 8 consecutive 13-residue windows (as explained in The predictor section). This average score is assigned to the first residue of the 20-mer and thus the graph starts at position 1 of the protein sequence.

The region shown in pink in this graph (the NoLS candidate signal region, which highlights scores above 0.8) represents the range of scores with which a 20 residue segment is predicted to be a NoLS.

Back to top

Contact Details

If you have any problems or queries with this website, please send an email to

Back to top