A

C

R

D

B


Citation: Huang et al., doi:10.1093/nar/gkaa857 2020
AcrFinder

Browse AcrDB by Source Database:


Taxonomy
Search
CRISPR Cas
419 Archaea 5980 Bacteria 2044 Virus

Introduction to AcrDB

AcrDB is a comprehensive database of computationally predicted anti-CRISPR (Acr) and Acr-associated (Aca) operons. Similar online databases include Anti-CRISPRdb, CRISPRminer, and AcrCatalog. Compared to these previous databases, AcrDB has the following unique features and data:

  1. It is a genome-scale database with the largest collection of data (39,799 Acr-Aca operons containing Aca or Acr homologs);
  2. It offers a user-friendly web interface with various functions for browsing, graphically viewing, searching, and batch downloading Acr-Aca operons;
  3. It focuses on the genomic context of Acr and Aca candidates instead of individual Acr protein families;
  4. It collects data with three independent programs (AcrFinder, AcRanker, and PaCRISPR) each having a unique data mining algorithm for cross-validation.

New Data and Features of AcrDB

In this update, we have significantly expanded AcrDB by including Acr operons (AOs) from human viromes and focusing on their predicted 3D structures. The process involved the following steps:

  1. Operon Prediction: We used AOminer to predict operons likely containing Acrs.
  2. 3D Structure Prediction: Proteins within putative Acr operons were subjected to 3D structure predictions using AlphaFold2.
  3. Filtering Predicted Structures: Predicted structures were filtered by comparing them to 122 experimentally verified Acrs using TM-Vec and Foldseek, and further validated by the latest machine learning-based tool, AcrPred.

The updated AcrDB includes the following new data compared to the previous version and existing resources:

  1. Predicted Acrs from the three largest human gut virome databases—GPD, GVD, MGV and the NCBI phage genome database INPHARED.
  2. 3D structures of 122 experimentally characterized Acr proteins reported in recent studies and predicted by ourselves.
  3. AlphaFold2 predicted 3D structures of 68,848 candidate Acrs with structural similarity (template modeling score or TM-score ≥ 0.6) to known Acrs supported by TM-Vec, 2,561 supported by Foldseek, and 8,649 by AcrPred18.
  4. A structural similarity search function to allow users submit new sequences and structures to search against 3D structures of the 122 known Acrs