How to use CE-BLAST Parameter Setting Available Tools Retrieve Files Visualization Announcement Data Citation Contact Us

How to use CE-BLAST

The accepted files of CE-BLAST currently contains antigen structures in PDB format (including crystal structures and modeled structures) and HA1 sequences of influenza virus (H1N1 and H3N2) in FASTA format. The work flow covers several steps. The next step will pop out when the previous step is finished. Please try to preprocess your files if error messages received.
In the section of "search against built-in database", the antigenic similarity score will be calculated between the query and structures in built-in database. In the database of 559 conformational epitopes, the fingerprints have been pre-calculated. When user specifies a self-defined epitope, CE-BLAST will calculate the fingerprints for query epitope, and then search the fingerprints database of known conformational epitopes. In the case of influenza virus, per-defined epitopes were proposed and user can defined new epitope areas according to their knowledge. If the input epitope positions are different from pre-defined ones, CE-BLAST will calculate new fingerprints for all the representative HA structures in our database based on the new epitope positions user defined. The whole workflow of CE-BLAST is illustrated as below:

After chose searching type between "search against built-in database" and "search against custom-defined database" in Home page. Steps for different uploading files are shown as follows:

1. Antigen Structures

Step 1). Uploading files

You can add new file to search multiple queries against built-in database; also, if you chose "search against custom-defined database" in Home page, you can add new file to define your own database for searching.

Step 2). Choosing Search Database

If you chose "search against built-in database", you can choose the database in this step, and specify the antigen family to search against.

Step 3). Specifying the epitope residues

Users can specify the epitope residues in antigen structures by entering the assigned residue numbers in the blank. For antigen structure, if no residues are specified, all residues in the structure will be treated as epitope residues.

View 3D structure can help you specify the epitope residues.

Step 4). Calculation

2. Influenza Sequences

Step 1). Uploading files

If you chose "search against custom-defined database" in Home page, you can upload multiple sequences file in one FASTA file to define your own database for searching.

Step 2). Choose representative structures of HA1

We have highly similar sequences of HA1 in our built-in database. After uploading HA1 sequences of influenza, CE-BLAST will recommend top 5 similar HA sequences according to sequence identities, and you can choose one to represent your input.

Step 3). Specifying the epitope residues

For influenza, user can chose known epitope types or specify user-defined epitope residues. Here, we provide two default Epitope Types which contain 16(Liao, et al., Bioinformatics models for predicting antigenic variants of influenza A/H3N2 virus. Bioinformatics) positions and 44(Smith, et al., Mapping the antigenic and genetic evolution of influenza virus. Science) positions respectively. Also, you can also define your own epitope residues by chosing "user_defined".

For influenza HA1 sequence, as the position numbers may shift after alignment, please double check your epitope residues from alignment against standard sequence.

3. Modeled Structures

For modeled structures of HA protein for influenza virus or E protein for DENV or ZIKV, user can model the structure and input it from "antigen structure" track. From there, you can choose predefined epitope position (for HA) or self-define your onw epitope positions. User can search against built-in database or searching within self-defined dataset.

Parameter Setting

1. Cutoff of Similarity Score

The default setting has been generalized for satisfying results, and user can set different thresholds for personal needs. Under a lower threshold, more epitopes will be selected and will lead to the increasing of true positive rata and simultaneously increasing the false positive rate as well. For CE-BLAST, 3 levels of similarity score were presented. Similarity score > 0.9 usually indicates highly similar epitopes within intra-subtypes of pathogens, 0.7 < similarity score < 0.9 usually indicates similar epitope areas within inter-subtypes, 0.6 < similarity score < 0.7 usually indicates epitopes with overlapping areas and similarity score < 0.6 usually indicates different epitopes. For intra- mutation analysis such as HA1 protein of influenza within specific subtypes (such as H3N2), the cutoff of similarity score was set as 0.9.
FFor inter- subtype analysis such as E protein of different DENV serotypes, the cutoff of similarity score was set as 0.7.

2. Parameters for Algorithm

For algorithm parameters, "horizontal" (h) and "vertical" (v) mean the resolution of spin-image, while "granularity" (c) means searching radius in seed grow. The default settings have been generalized for satisfying results. It should be noticed that high resolution and granularity will lead to high computational complexity, however, may not always lead the increasing of performance.

Available Tools

1. Re-label multiple chains into single one

As CE-BLAST only accept protein structure with one single chain, epitopes cover two chains need to be merged into one chain with unique residue numbers.

After uploading your file with multiple chains, this tool will automatically check chain names.

You can select multiple chains need to be merged, and input the new chain name in the blank.

2. Re-number residues with non-numeric identifiers

Please note that PDB files may contains residue ids with non-numeric identifiers such as 100, 100A, 100B; this may sometimes causing trouble for calculation. You can make all residue ids in PDB file with one unique numerical ids by using this tool.

3. Epitope identification

This tool can help you identify epitope residues in antigen-antibody complexes.

After specifying both antigen and antibody chains, you can also select different cutoff of distance. All antigen residues with least atom distance towards antibody residues within cutoff will be extracted as epitope residues.

Retrieve Files

After receiving the result report file, you can retrieve the hits epitope files here by uploading a list of hit ids.


You can visualize your structures by using plug-in "Jmol", your selected residues will be marked in gold for visualization.


The sequence alignments are calculated by BLAST 2.2.25+.
BLAST: Basic Local Alignment Search Tool

Jmol 14.4.1 has been implemented into CE-BLAST.
Jmol: an open-source Java viewer for chemical structure in 3D

Data Citation

Contact Us

1239 Siping Road, Shanghai, China