Protein Domains and Macromolecular Structures
What's New
   Conserved Domains   3D Macromolecular Structures 
  Conserved Domains and Protein Classification back to top  

[20 SEP 2022]  A new version of the Conserved Domain Database has been released. Version 3.20 contains 1,614 new or updated NCBI-curated domains and now mirrors Pfam version 34 as well as new models from the NCBIfam collection. Fine-grained classifications of the [(+)ssRNA] virus RNA-dependent RNA polymerase catalytic domain, RING-finger/U-box, dimerization/docking domains of the cAMP-dependent protein kinase regulatory subunit, and Galactose/rhamnose-binding lectin domain superfamilies have been added, amongst many other new models. You can access CDD at and find updated content on the CDD ftp site at

[8 MAR 2021]  A new version of the Conserved Domain Database has been released. Version 3.19 contains 3,148 new or updated NCBI-curated domains and now mirrors Pfam version 33.1 as well as models from the NCBIfam collection. Fine-grained classifications of the immunoglobulin, RRM, cytochrome P450, 7-transmembrane GPCRs, KH, calponin homology and C1 domain superfamilies have also been added. With this release, CDD introduces model-specific word-score thresholds for the RPS-BLAST heuristics. These are included in the position-specific score matrices (PSSMs) and are used by database formatting software when constructing word lookup tables for the BLAST heuristics stage. The current implementation results in a 3-fold speedup of RPS-BLAST searches and misses annotation for about 0.6% of query proteins, mostly at the borderline of significance. You can access CDD at and find updated content on the CDD ftp site at

[30 APR 2020]   New viral protein domain models for annotation of coronaviruses are now available. NLM's Conserved Domain Database (CDD) has expanded its scope to now include 153 new viral protein domain family models for the annotation of coronaviruses, including models such as for the S1 subunit of coronavirus Spike proteins (cd21527), the nucleocapsid (N) protein of coronavirus (cd21595), and the coronavirus RNA-dependent RNA polymerase (cd21530). Each curated domain model consists of a multiple sequence alignment illustrating conserved amino acids and may define conserved sequence features that have been confirmed experimentally, plus links to relevant publications. When available, the domain models include 3D structures with links to interactive 3D views and interacting partners.

A tabular summary of SARS-CoV-2 gene products along with links to matching conserved domain models and representative 3D protein structures is available here.

[26 MAR 2020]  A new version of the Conserved Domain Database (CDD) has been released. Version 3.18 contains 2,128 new or updated NCBI-curated domains and now mirrors Pfam version 32 as well as additional models from the NCBIfam collection. Fine-grained classifications of the cupin and PBP1 superfamilies have also been added. You can access CDD from and find updated content on the CDD FTP site at Database statistics, showing the number of domain models from each source database, are provided on the CDD News page.

[08 JAN 2020]  The print version of CDD/SPARCLE: the conserved domain database in 2020  by Lu S et al. is now available in the Nucleic Acids Research database issue (PubMed ID 31777944; full text at Oxford Academic; full text in PubMed Central).

[28 NOV 2019]  CDD/SPARCLE: the conserved domain database in 2020  An article by Lu S et al., about the NCBI Conserved Domain Database became available in Nucleic Acids Research as an e-publication ahead of print (PubMed ID 31777944): "As NLM's Conserved Domain Database (CDD) enters its 20th year of operations as a publicly available resource, CDD curation staff continues to develop hierarchical classifications of widely distributed protein domain families, and to record conserved sites associated with molecular function, so that they can be mapped onto user queries in support of hypothesis-driven biomolecular research..." (read more)

   More News   Subscribe to Mailing List 
  3D Macromolecular Structures back to top  

[17 MAY 2021] Batch Process of 3D Structures Using iCn3D. With the recent release of "icn3d" package at npm and iCn3D 3.0, you can write Node.js scripts to call functions in icn3d. These scripts can be run in the command line to process a list of 3D structures to get annotations. For example this script calculates the change in interactions due to a mutation. The classes and functions are listed at iCn3D Doc. To find icn3d functions, you can first generate a custom view in iCn3D interactively. For example, you can click the menu "Analysis > Mutation" to see the "Interactions" of the mutations "6M0J_E_501_Y". Then you can click the menu "File > Share Link" to see the commands in the section "Original URL with commands". Most commands are processed with functions in the file "applyCommand.js". Other commands with asynchronized retrieval are processed in the file "loadScript.js". Since the JavaScript code of iCn3D was upgraded from ES5 to ES6, the embedding of iCn3D has some changes.


[30 NOV 2020]  iCn3D 2.22.0 is now available on NCBI web servers and from GitHub ( With the recent release of iCn3D 2.22.0, users can now visualize consequences of non-synonymous sequence variation and alternate between 3D views for wild type and substituted side-chains for selected SNPs, as in this example. In the "Sequences & Annotations" window, which is accessible in the menu "Analysis > View Sequences & Annotations", users can click the tab "Details" and the checkbox "SNPs" or "ClinVar" to see known sequence variation, if available. The mouseover on SNPs/ClinVar residues with 3D-stucture shows three buttons: "3D with scap", "Interactions", and "PDB", which allow users to alternate between the wild type and mutant structures, alternate between their interactions, or download their coordinates. When users click "Interactions", the 2D interaction network window also pops up to show the change of interactions, which are shown in different colors. The coordinates of the mutant are predicted dynamically with the scap web service, which is converted from the scap program ( The web service takes the coordinates of the residues within 10 angstrom from the SNP, uses rotamer libraries to predict the side chain conformation, and outputs the coordinates of the mutant and the neighboring residues. Another new feature dynamically computes and displays a selected protein's internal pseudo-symmetry, as in this example. The symmetry is calculated with the symd web service, which is converted from the SymD program (


[07 OCT 2020]  iCn3D 2.20.0 is now available on NCBI web servers and from GitHub ( Users can now view the electrostatic potential map for any subset of 3D structures within 30,000 atoms. The potential is calculated using the DelPhi program by solving linear Poisson-Boltzmann equation. Users can show the potential on surface or show equipotential map. The potential map shows the effect of charges on molecular interactions qualitatively. This example shows the electrostatic potential for the binding of Gleevec to Abl2. The ligand shows the -25 mV (red) and +25 mV (blue) equipotential map with a grid size 65, salt concentration 0.15 M, and pH 7. The protein shows the surface potential with a gradient from -75 mV (red) to +75 mV (blue). This new feature can be accessed from the menu "Analysis > DelPhi Potential". Users can also download the PQR file format with assigned partial charges.

Other recently introduced features include:
More features are listed and described at

[22 JUL 2020]  iCn3D 2.18.0 is now available on NCBI web servers and from GitHub ( Users can view SNPs of the 2019-nCov structures in the "Sequences & Annotations" window. Users can also access the 2D Interaction Network and 2D Interaction Map for protein-ligand or protein-protein interaction by selecting the menu "View > H-Bonds & Interactions". Another new feature is "icn3dpy", a Jupyter Notebook widget of iCn3D, which enables data scientists to use iCn3D in a Jupyter environment. Users can view a predefined display or create a custom view directly in the Jupyter Notebook. As an example the SNPs of the 2019-nCov spike (S) glycoprotein were viewed using a Jupyter Notebook and saved here as HTML.


[21 APR 2020]  iCn3D 2.15.0 is now available on NCBI web servers and from GitHub ( To view the updated web application, retrieve any structure of interest from the Molecular Modeling Database (MMDB), open its structure summary page, and click the button for "full-featured 3D viewer" in the molecular graphic. For example, retrieve structures that contain the term "SARS-COV-2", click on a structure of interest to open its summary page, then follow the link for "full-featured 3D viewer." Alternatively, open iCn3D and use the "File" menu to retrieve a structure by its ID (for example, enter 2DD8 to load the crystal structure of Sars-Cov Spike Receptor-Binding Domain Complexed With Neutralizing Antibody) or to open a structure file on your local computer. A gallery with live examples demonstrates a variety of ways in which iCn3D can be used to view and analyze structures. The iCn3D Web APIs document describes how to use the iCn3D structure viewer in your own web page.

The gallery with live examples now includes COVID-19-related structures:
New features in this iCn3D release include the ability to:
  • Apply custom colors to individual residues by using the iCn3D menu option for "Color > Residue > Custom."
  • Add "Custom Color" for any chain in the "Sequences & Annotations" window.
  • Add multiple sequence alignments as tracks when clicking "Add Track" in the "Sequences & Annotations" window.
  • Show the same structure "Side by Side" in two views in the "View" menu. Each view has the same orientation, but can have an independent 3D display. One example is shown in the gallery, where each view shows one of the aligned structures.
  • Realign two structures or two chains by using the iCn3D menu option for "File > Realign," if you want to view an alignment that is different than the one generated by VAST+ or by dynamic chain alignment. Specifically, the "Realign" option enables you to select residues of interest and to realign the structures or chains on those residues.
    For example, some chains, such as 6ACK_C and 6M0J_E don't align well using dynamic alignment. Using the "Realign" function, you can select those two chains, which are then aligned based on their sequence data. The 3D coordinates of the residues in the sequence alignment are then used to generate a new structure alignment.
  • The Change Log section of the iCn3D Web API help document lists additional enhancements that have been made to iCn3D since its original release.

[20 JUN 2019]  Bioinformatics publication describes features and applications of iCn3D, NCBI's web-based 3D viewer, and provides examples of its use for interactive structural analysis:

  • Wang J, Youkharibache P, Zhang D, Lanczycki CJ, Geer RC, Madej T, Phan L, Ward M, Lu S, Marchler GH, Wang Y, Bryant SH, Geer LY, Marchler-Bauer A. iCn3D, a Web-based 3D Viewer for Sharing 1D/2D/3D Representations of Biomolecular Structures. Bioinformatics. 2019 June 20; pii: btz502. doi: 10.1093/bioinformatics/btz502. [Epub ahead of print] [PubMed PMID: 31218344] [Full Text at Oxford Academic]

[01 APR 2019]  Multi-character identifiers for sequences with 3D structures from the PDB.  Advances in experimental methods permit the three-dimensional (3D) structure determination of ever larger biomolecular assemblies. To accommodate this, the RCSB Protein Data Bank (PDB) relaxed their constraint that individual biopolymers (proteins and nucleotide sequences) in 3D macromolecular structure records are labeled by a single character. Specifically, a biopolymer identifier may now be up to four characters long.

  • This change has required NCBI to revise how it treats PDB-derived sequence records.
  • The initial and most visible impact will appear in the release of BLASTDBv5 (see BLAST announcement), which will support multi-character identifiers for PDB-derived sequences. But any software, from NCBI or created elsewhere, that assumes a one-character PDB chain identifier will be affected.
  • These changes will also impact the Molecular Modeling Database (MMDB), which is derived from PDB, and the related structure services. More details about how NCBI structure services are affected will be provided on the MMDB News page once multi-character identifiers are fully implemented in MMDB.
   More News   Subscribe to Mailing List 
Revised 21 September 2022