Session unregistered

Your session timed out and has been unregistered. You will still be able to access data on pages that you have already submitted, but you will not be able to submit any new requests to the Prism server without reloading the page to register a new session.

Click here to reload Prism now.

Unsupported browser

You are using an old or unsupported browser that has not been tested with this version of Prism. Prism relies on new browser technologies to display genome search results. For optimal performance, we recommend using Prism with the latest version of Google Chrome.


Prediction informatics for secondary metabolomes:
search a genome for genetically encoded natural products

PRISM's combinatorialization process can lead to the generation of a large number of structures for some biosynthetic gene clusters. By default, PRISM will limit the number of putative structures to be generated to 50.


Select the method used to identify putative open reading frames within your submitted sequence. (Note that Prodigal requires a minimum sequence size of 20,000 nucleotides.)

Optionally provide a file containing the coordinates of open reading frames within the input sequence in GFF3 format.

Specify the maximum length between open reading frames, in base pairs, to consider them part of the same biosynthetic cluster.


Enable or disable search for specific families of biosynthetic domains, each of which extends the runtime of a PRISM search.

Show advanced settings

Open saved results

Load saved PRISM output for a microbial genome in JSON format.

Select a saved PRISM results file in JSON format.


December 31, 2016
PRISM 3 is online now. We're rolling out some big changes with PRISM 3. The biggest change is that the chemical structure prediction engine within PRISM has been completely rewritten. Instead of modelling natural product structures as linear permutations of monomers, PRISM now models them as chemical graphs. This allowed us to extend structure prediction to four new natural product classes—aminocoumarins, antimetabolites, bisindoles, and phosphonate-containing natural products. We've also added cluster detection, but not structure prediction, for eleven new classes (aryl polyenes, butyrolactones, ectoines, furans, homoserine lactones, ladderanes, melanins, phenazines, phosphoglycolipids, resorcinols, and stilbenes). Finally, PRISM 3 also features improved sequence and ORF detection, a more user-friendly interface, a more detailed Help page with sample output, and improved JSON and HTML output.

October 4, 2016
You may notice some changes to PRISM associated with the public release of the PRISM 2.1.5 web application. We’ve rewritten the version of PRISM deployed at to run over a 300-core server grid. This means PRISM will no longer give detailed updates on the status of your job every second. However, jobs should run much faster, especially during times when PRISM is in high demand. We hope you like it.

In other news, the RiPP-PRISM paper is now published. RiPP-PRISM is integrated into the publicly available PRISM web application: users have the option of enabling or disabling RiPP cluster analysis and prediction.


Step 1: upload a sequence file

PRISM accepts DNA sequence input in FASTA or GenBank format. When input sequences contain over 1,000 contigs, only the 1,000 largest will be considered in order to minimize server load. Only contigs above 500 nt will be considered. Upload a nucleotide sequence by selecting “Choose file.” Alternatively, a sample input file can be loaded by selecting “Load sample input,” or annotated sample output can be opened by selecting “Open annotated output.”

Step 2: configure advanced options (optional)

PRISM makes several advanced options available to users with specific needs. In general, users do not need to configure these parameters.

Structure limit: For clusters with a large number of potential open reading frame permutations, or tailoring enzymes, hundreds or even thousands of combinatorial structures may be generated, significantly extending the runtime of a PRISM search. For this reason, we suggest limiting the maximum size of each cluster's combinatorial structure library. A pseudorandom selection of combinatorial plans will have their corresponding structures generated.

Open reading frame prediction: By default, PRISM will locate all possible coding sequences between start and end codons. Additionally, PRISM uses Prodigal to predict prokaryotic genes. When potential coding sequences overlap with Prodigal-predicted genes, the former are removed. Either option can be disabled, or PRISM can simply read ORFs directly from an annotated GenBank sequence file.

GTF file: A fourth option for ORF detection is to input the coordinates of open reading frames within the sequence file in GTF format.

Window: Increase or decrease the clustering window that PRISM's greedy algorithm uses for determining the boundaries of each biosynthetic gene cluster.

Optional searches: PRISM is capable of predicting the chemical structures of eleven classes of secondary metabolites, and identifying biosynthetic gene clusters for an additional twelve, as well as detecting resistance genes within biosynthetic gene clusters. Each search extends the runtime of a PRISM search, and consequently the option is made available to disable one or more of them.

Step 4: submit

Once the Submit button is clicked, PRISM will automatically search the input sequence for genetically encoded natural products. The link at which your results will appear, and the status of your job on the PRISM server grid, is provided.

Step 5: understand your results

In order to facilitate interpretation of PRISM results, we have created an annotated output page, which explains how the HTML output corresponds to various functionalities executed within PRISM.


PRISM is a project of the Nathan Magarvey Lab at McMaster University. We aim to provide bio- and chemoinformatic solutions for genomic natural product discovery.

PRISM is an algorithm for the prediction of genetically encoded natural product structures based on microbial genomes. It takes a microbial genome sequence as input, identifies biosynthetic gene clusters, and generates combinatorial libraries of structure predictions.

PRISM is freely available as an online service for the research community. PRISM implements the Chemistry Development Kit1, hmmer2, BLAST3, BioJava4, and FIMO5.

Citing PRISM

Skinnider, M.A., Dejong, C.A., Rees, P.N., Johnston, C.W., Li, H., Webster, A.L.H., Wyatt, M.A., and Magarvey, N.A. (2015). Genomes to natural products prediction informatics for secondary metabolomes (PRISM). Nucleic Acids Research, 43, 9645-9662. doi: 10.1093/nar/gkv1012.

Skinnider, M.A., Johnston, C.W., Edgar, R.E., Dejong, C.A., Merwin, N.J., Rees, P.N., and Magarvey, N.A. (2016) Genomic charting of ribosomally synthesized natural product chemical space facilitates targeted mining. Proceedings of the National Academy of Sciences, 113, E6343-E6351. doi: 10.1073/pnas.1609014113.


1 Steinbeck, C., Han, Y., Kuhn, S., Horlacher, O., Luttmann, E. and Willighagen, E. (2003) The Chemistry Development Kit (CDK): an open-source Java library for Chemo- and Bioinformatics. J. Chem. Inf. Comput. Sci., 43, 493-500.

2 Finn, R.D., Clements, J. and Eddy, S.R. (2011) HMMER web server: interactive sequence similarity searching. Nucleic Acids Res., 39, W29-37.

3 Altschul, S.F., Gish, W., Miller, W., Myers, E.W. and Lipman, D.J. (1990) Basic local alignment search tool. J. Mol. Biol., 215, 403-410.

4 Prlic, A., Yates, A., Bliven, S.E., Rose, P.W., Jacobsen, J., Troshin, P.V., Chapman, M., Gao, J., Koh, C.H., Foisy, S. et al. (2012) BioJava: an open-source framework for bioinformatics in 2012. Bioinformatics, 28, 2693-2695.

5 Grant, C.E., Bailey, T.L., and Noble, W.S. (2012) FIMO: scanning for occurrences of a given motif. Bioinformatics, 27, 1017-1018.