A2CA: Amino Acid Cluster Analysis
Web version …
Protein function emerges from a highly coupled three-dimensional scaffold in which substitutions can influence interactions far beyond an active or binding site. This complexity makes it difficult to distinguish residues that merely vary from positions constrained by structure, chemistry, or compensatory evolution, complicating both mechanistic interpretation and rational protein engineering. A2CA integrates multiple sequence alignments, phylogenetic trees, amino-acid physicochemical properties, optional protein structures, and pairwise coevolution analyses to place residue-level variation into an evolutionary context and help prioritize positions for experimental investigation.
Contact: Dr. Daniel Eggerichs
How would you like to start?
Choose the entry point that matches your available data. A2CA can retrieve homologous proteins from a single sequence or protein structure with NCBI BLAST, generate an alignment and phylogeny from unaligned FASTA sequences, or start directly from a prepared alignment and tree.
Resources
The original R/Shiny implementation of A2CA is archived as: D. Eggerichs, D. Tischler; Science Data Bank (2023), DOI: www.doi.org/10.57760/sciencedb.09549.
Source code, version history, and future releases are maintained in the A2CA GitHub repository.
Data privacy
Most A2CA analysis data and saved sessions are handled locally in your browser. In the BLAST workflow, only the selected protein query sequence is forwarded to NCBI; uploaded structure files themselves are not sent to NCBI. Unaligned FASTA submitted for alignment and tree construction is processed on the A2CA server by MAFFT and FastTree and is not forwarded to a third-party alignment or tree service. Identifiers used to fetch NCBI Protein or RCSB PDB records are sent to the corresponding external service.
Software and services
- MAFFT, executed on the A2CA server using the automatic strategy, for multiple-sequence alignment.
- FastTree 2.1.11, executed on the A2CA server, for phylogenetic tree inference.
- 3Dmol.js 2.5.5 for interactive WebGL protein-structure visualization.
- RCSB Protein Data Bank for retrieval of structures by PDB identifier.
- NCBI BLAST Common URL API and NCBI Protein EFetch for homolog discovery and retrieval of cluster-representative protein sequences from a single query.
- The A2CA interface and analyses use browser-native HTML, CSS, JavaScript, SVG, the File API, and WebGL. A small hosted Python backend provides same-origin access to NCBI and RCSB services and executes MAFFT and FastTree.