The suite of programs developed by our group for accurately identifying protein secondary structures.
(Emap2Sec), protein nucleic acid (RNA/DNA) (Emap2Sec+), full atom 3D model MAINMAST or
segmentation of individual components (MainMastSeg), improved protein modeling from GAN-modified EM maps (EM-GAN), model quality assessment with (DAQ Score), model quality assessment results in (DAQ-Score Database), local model refinement with (DAQ-refine), real-time comparison and analysis of cryo-Electron Microscopy (EM) maps (EM-Surfer), and EM map superimposition (VESPER).

TOOLS and Databases

List of our cryo-EM Modeling Tools and Databases


Emap2sec uses deep learning to accurately identify secondary structures (alpha helices, beta sheets, coils/turn) in cryo-EM maps of medium to low resolution.


Emap2sec uses deep learning to accurately identify protein secondary structures and nucleic acid (RNA/DNA), in cryo-EM maps of medium to low resolution.


MAINchain Model trAcing using Spanning Tree from a EM map, directly traces main-chain connections and C-alpha positions by using Tree-Graph models.


Minimum Spanning Tree based EM map segmentation with Symmetry restraints. MAINMASTseg directly segment out individual components from Electron Microscopy (EM) density map.


EM-GAN is a computational tool, which enables capturing protein structure information from cryo-EM maps more effectively than raw maps. It is based on 3D deep learning. It is aimed to help protein structure modeling from cryo-EM maps.


DAQ-Score (Deep learning-based Amino acid-wise Quality assessment score) evaluates agreement of a protein model with detected amino acid residues and other structural features in a density map by deep learning.

DAQ-Score Database

DAQ-Score Database stores precomputed residue-wise local quality scores of protein models from cryo-EM maps.


DAQ-refine refines low-quality regions in a protein structure model by DAQ score and AlphaFold2.


A web-based tool for real-time comparison and analysis of Electron Microscopy (EM) density maps. It compares the shape of EM map isosurfaces, generated using author-recommended contour values.


VESPER uses a combination of mean shift algorithm and fast Fourier transform (FFT) to identify the best superimposition of two EM maps.