General site navigation
To submit a kinase sequence or view/search the kinase/substrate database, you must be logged in as a registered user of the site. When you are logged in a user menu becomes visible to you. From here you can always access your current predictions. You can view all of the other menu links without logging in, but the discussion forums will be read-only.
You can also navigate via tabs at the top of the page when you are analysing a kinase sequence. Note that if you return to the "Submit a kinase" link, your current kinase is cleared (i.e. "Your current prediction" will be blank). However, the "All your predictions" link will store all your data from a session for 24 hours.

1. Sequence entry
Once logged in, the "Submit a kinase" link becomes available. You can paste a sequence into the text area or upload a file from disk. In either case only one sequence is allowed and it must be in fasta format. For a sample sequence, click the "Example" button.
Click "Submit" when you are ready to analyse the sequence.

2. Kinase data
Predikin will analyse your sequence to determine the kinase type (Ser/Thr, CMGC or Tyr), KSD family, PANTHER family and SDRs. Each step of the analysis is displayed as it is completed.

Assuming that your sequence is error-free and is recognised as a protein kinase, you will be directed to the kinase data page. This page displays a summary for each kinase domain in your sequence, with links to the Kinase Sequence Database and PANTHER websites.

3. Choosing a matrix
Use the tabs at the top of the page to examine the SDR, KSD and PANTHER matrices. Note that for a tyrosine kinase, only the latter two matrices are generated.
Yellow cells indicate the highest frequency (or weight) in each row. There may be more than one yellow cell per row if several values are equal highest.
Red cells indicate that the total frequency for a row is zero. This can only occur using the SDR method. If this occurs, the SDR weight matrix is incomplete and should be used with caution or not at all.
For a Ser/Thr or CMGC kinase, the SDR matrix should normally be your first choice (unless incomplete). Matrices for which the row sum is small (e.g. < 10) are less reliable. Matrices in which many cells contain the same frequency are also unreliable - they are probably derived from a small number of identical phosphorylation sites.

4. Scoring a substrate
Once you're happy with your matrix, click the "Score substrates" tab. Here, you can paste or upload a putative substrate sequence, again in fasta format. Choose a scoring method (SDR, KSD or PANTHER) and scoring parameters. A minimum cutoff score of 60 is the recommended default. It is also recommended that you select "ignore sites in TM helix" and "ignore non-disordered sites", as the vast majority of known phosphorylation sites in our database satisfy these two criteria.
When ready, click the "Submit" button and wait a few seconds.

5. Substrate scores
Substrate scores are displayed as a table. The table columns are kinase ID, domain, substrate ID, position of S/T/Y, heptapeptide (-3 to +3), scoring method and score. No table rows means no scores above your specified cutoff.
The rows are sorted by score, highest to lowest. You can return to the scores for your current kinase at any time using the "Your current prediction" link in the user menu. You can also return repeatedly to the "Score substrates" tab and so input multiple substrates for a kinase.

6. Saving and exporting predictions
All of the predictions that you generate at the Predikin website are stored for 24 hours. In this way you can build up a dataset of results for multiple kinases and substrates. When you're finished, go to the "All your predictions" link in the user menu. Click "Export" to save your results in a CSV file that you can download and import to a spreadsheet or local database.