Input – Sequences

Paste your amino acide sequences in FASTA format such as:


where AC_1 and AC_2 are sequence IDs. Each sequence should be at least 5 a.a. long and maximally 250 a.a. or an error will be reported. Only 1-letter code for 20 standard amino acids is acceptable. If the number of sequences is large, you can upload them in a file (in FASTA format) by selecting the “Upload FASTA sequences” checkbox. If you want to screen a genome sequence to identify AMPs from that, use the “Upload Genome sequences” checkbox. In this case, only 1-letter code for nucleotides (ATCG) is acceptable. Genome sequences will be processed to extract the open reading frames (ORF) in all six-frame translation using the the user-selected codon table (standard by default). ORF sequences which include arbitrary residues will be discarded; sequences which are shorter than 5 residues and longer than 250 residues will also be ignored. Uploaded sequence files are limited to 50 MB.

Input – Methods

Select the antimicrobial activities that you are interested to predict:

  • AmPEP for general AMP prediction of 5-250 AA sequences.
  • Deep-AmPEP30 for short AMP prediction of 5-30 AA sequences (According to our benchmark test, Deep-AmPEP30 is about 18% more accurate in short AMP prediction than AmPEP). This is the only method currently available for genome screening.
  • AcPEP (trial version) for anticancer prediction, still under development.

Input – Job Description (Optional)

A short description can be supplied together with a job submission. The description will be displayed in your job list.

Input – Create Account (Necessary)

If you provide an email to your submitted job, you can easily retrieve all your past job results by giving your email. Note that we will not send you any emails about your job run.

Input – Submit

Click the submit button when ready. A job ID will be given to your submitted job; a status page will be displayed with automatically refreshing function to check for the running status of your job. As soon as the run is finished, the result will be displayed in the same window. Please bookmark the status page before leaving the window, so that you can check back the result later. Or if you have provided an email during job submission, you can enter your email at the Input page to login and retrieve all your past job results.


The prediction result of all submitted sequences will be displayed in a single page. All columns can be sorted as user’s need, and all results are downloadable in a CSV file.

The Classification tag shows 1 or 0 predicted by each selected method, where 1 for positive and 0 for negative. In case there are invalid sequences, such as violating the sequence length required by the method, they are classified as -1. The last column shows the sum of the prediction results for each sequence. A classification threshold of 0.5 is used by default to color code the result table.

The Prediction Score tag shows the predicted score returned by each method, with a value close to 1 being positive and close to 0 being negative. Again, invalid sequences are given -1. The last column shows the product of all predicted scores for each sequence. A classification threshold of 0.5 is used by default to color code the result table.

For genome screening, only the top-scoring 1000 sequences will be displayed. Prediction result of all valid sequences is available in the CSV file.


If you have further questions, please contact:


  • Pratiti Bhadra
  • Jielu Yan
  • Edison Un (web developer)


  • Shirley W. I. Siu