Peptides Bioactive Functional Annotation

Introducing PepAnno

PepAnno combines sequence-based and structure-based computational models to predict multiple peptide bioactivities simultaneously, covering seven major functional categories, including antimicrobial and anticancer properties. In addition to analytical capabilities, PepAnno incorporates a curated repository of peptide-related computational resources, providing users with a centralized hub for data access and comparative research. By unifying functional prediction, structural analysis, and access to curated peptide-related resources, PepAnno allows researchers to rapidly explore peptide properties, compare functional hypotheses, and generate biologically meaningful insights.

Go Prediction See Example

The overview illustration of PepAnno's framework

Structure-Aware Multi-view Deep Learning Framework

To achieve accurate identification and functional annotation of bioactive peptides across diverse categories, we propose a novel Structure-Aware Multi-view Geometric Deep Learning Framework. This framework synergistically integrates three core components: (1) a multi-view data representation module that constructs heterogeneous graphs from sequence and structural information; (2) a dual-stream neural architecture utilizing cross-modal attention for deep feature fusion; and (3) a strict hierarchical transfer learning strategy designed to ensure robust generalization on small-sample datasets.

The help to use PepAnno

Predict the peptides you're interested in.

First put the peptide sequences you are interested in into the prediction area, make sure they are in fasta format. If you don't know what to do, click the Example button to get the reference input and reference selection, and you can also see an example of the final result through the see example result button. You can compute a peptide profile by amino acid scale. For various categories of scale, you can click to expand and select the scale you are interested in. Hydropathicity Scale of Hydropath Kyte & Doolittle and Transmembrane tendency are selected by default. You can change the window size, which is the length of the interval to use for the profile computation. The default is 9.

Parameter selection of bioactive function & Other optional input.

In the prediction interface, users may select from 54 predefined amino acid scales (e.g., hydropathicity and transmembrane tendency scales) for profile computation and adjust the sliding window size as needed. Upon submission, PepAnno automatically executes the selected prediction workflows. For optional model-based predictions, users can specify one or more of the seven supported bioactive peptide functions and either select individual predictive models or adopt the default parameter settings.

Result Visualization - General Information

For the basic information of peptides, we provide a table that allows you to switch between input sequence IDs to display an overview of the data. The table includes details such as the sequence, sequence length, and the predicted biological activity. Additionally, based on the sequence composition, we present two visualizations: a bar chart showing the amino acid composition and a line chart illustrating the amino acid percentage. You can select individual sequences from the legend for focused analysis according to your interests. Icons are also available for selecting and displaying specific information relevant to your research.

Result Visualization - Physical-chemical Information

For the physical-chemical properties of peptides, we provide a table that allows you to switch between input sequence IDs to display an overview of the data. The table includes details such as the sequence, molecular weight, aromaticity, instability index, and isoelectric point. Additionally, we offer line charts illustrating 19 different physical-chemical property types based on sequence composition. You can select individual sequences from the legend for focused analysis according to your interests. Icons are also available for selecting and displaying specific information relevant to your research.

Result Visualization - Structural information

For the structural features of peptides, we provide a table that allows you to switch between input sequence IDs to display an overview of the data. The table includes details such as alpha helix, beta turn, and beta sheet information. Additionally, we use bar charts to show the likelihood of each amino acid position belonging to a specific secondary structure type for each sequence. Tertiary structures are predicted using ESMfold, with visualizations and downloadable PDB links provided. You can select individual sequences from the legend for focused analysis according to your interests.

Result Visualization - Get the bioactive function

Bioactive Function results are shown using a global summary table and radar charts depicting predicted activities for all input peptides. For individual peptides, detailed prediction scores and interpretable sequence-level attention profiles are provided. For predictions generated using optional models, an additional integrated table summarizes and compares scores across the selected methods