AntigenMap is an online resource developed specifically for antigenic cartography construction using the data from immunological experiments, such as hemagglutination inhibition, microneutralization, and Enzyme-linked immunosorbent (ELISA) assays.
Antigenic variation is a common strategy for an infectious pathogen to evade its host immunological response. This infectious pathogen can alter its surface protein(s) so that it will not be recognized by the specific antibody from the host. As a result, this pathogen can re-infect the same host. Antigenic variation can be either antigenic drift from accumulating mutations or antigenic shift from exchanging surface proteins from two different strains. For influenza A virus, antigenic drift may lead to seasonal epidemics and antigenic shift to pandemics in human.
Hemagglutination inhibition (HI), microneutralization (MN), and Enzyme-linked immunosorbent (ELISA) assays are three commonly used techniques for antigenic variation characterization.
Analyses of these immunological data are not trivial for the following reasons: (1) these assays are very crude and much noises are present in these datasets; (2) it is not uncommon that there are empty entries in the resulting HI, MN, or ELISA tables; (3) the data distribution in these datasets are unique, and those unobserved values are not randomly distributed in the tables, especially for the dataset which contains viruses with large isolation time interval. The details about these challenges are available in [1].
Antigenic map is an analog of geographic cartography in presenting the comparing antigens onto two or three dimensional map. Antigenic map has been shown to effectively detect antigenic drift events, thus it is useful in antigenic characterization and vaccine strain selection. Influenza antigenic map has becoming a standard tool in influenza surveillance of the World Health Organization (WHO) influenza reference laboratories.
The temporal MC-MDS method integrates a low rank matrix completion algorithm and a multidimensional scaling method with temporal modeling. Briefly, a low rank matrix completion algorithm is applied to fill in the entries of the HI matrix by assuming each antigen i can be embedded into the r dimensional space as ui and antiserum j can be embedded into the r dimensional space as vj. Then a temporal MDS algorithm is utilized to map the antigens (or similarly, antibodies) into a two dimensional space for visualization. The mathematical details of this algorithm was described in [1]. MC-MDS was shown to effectively minimize the biases from two major noises in the immunological data, missing values and low reactors. The temporal model was shown to be especially useful for antigenic map construction for the immunological datasets spanning a long time period. For instance, in influenza virus, this time span can be 12 years or more.
AntigenMap provides users a choice to construct antigenic cartography in either 2D or 3D format. A 3D antigenic map can overcome the limitation of 2D map by providing an additional dimension in the visualization. This is especially critical for the complicated data, such as immunological data. 3D display of antigenic cartography data allows for higher accuracy, especially when combined with the temporal model. Below is a image showing the difference.
Acmacs is a webserver developed for antigenic cartography construction. Currently it is still in testing time period.
ATIVS is a webserver for analyzing serological data of all influenza viruses and hemagglutinin sequence data of human influenza A/H3N2 viruses so as to generate antigenic maps for influenza surveillance and vaccine strain selection.