TM-align: a protein structure alignment algorithm based on the TM-score

By Yang Zhang, Jeffrey Skolnick. 2005

3,652 citations recorded by the source corpus.

We have developed TM-align, a new algorithm to identify the best structural alignment between protein pairs that combines the TM-score rotation matrix and Dynamic Programming (DP). The algorithm is ∼4 times faster than CE and 20 times faster than DALI and SAL. On average, the resulting structure alignments have higher accuracy and coverage than those provided by these most often-used methods. TM-align is applied to an all-against-all structure comparison of 10 515 representative protein chains from the Protein Data Bank (PDB) with a sequence identity cutoff <95%: 1996 distinct folds are found when a TM-score threshold of 0.5 is used. We also use TM-align to match the models predicted by TASSER for solved non-homologous proteins in PDB. For both folded and misfolded models, TM-align can almost always find close structural analogs, with an average root mean square deviation, RMSD, of 3 Å and 87% alignment coverage. Nevertheless, there exists a significant correlation between the correctness of the predicted structure and the structural similarity of the model to the other proteins in the PDB. This correlation could be used to assist in model selection in blind protein structure predictions. The TM-align program is freely downloadable at .

Subfields: Materials Chemistry, Molecular Biology, Computational Biology & Bioinformatics

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Materials Chemistry

In Molecular Biology

In Computational Biology & Bioinformatics

Citations in the corpus

Robust deep learning based protein sequence design using ProteinMPNN (2022, neutral): “We split those clusters randomly into three groups for training (23,358), validation (1,464), and testing (1,539), ensuring that none of the chains from the target chain or chains from the biounits of the target chain would be in the other two groups.”
End-to-end protein–ligand complex structure generation with diffusion-based generative models (2022, methodology): “We used TM-align [43] for structural alignment with experimental structures and used the resulting TM-score to evaluate protein conformations.”
State-specific protein–ligand complex structure prediction with a multiscale deep generative model (2022, positive): “The holo structures are obtained from the intersection between the filtered set and the annotated protein–ligand complex database BioLip 71 41 and distinct bound ligands can be found from the PDB.”
Accurate RNA 3D structure prediction using a language model-based deep learning approach (2022, positive): “Assessed using the template modeling (TM) score 31 , RhoFold+ achieved an average of 0.57 (Supplementary Table 5), higher than the scores of other top performers (0.41 and 0.44).”
Sampling alternative conformational states of transporters and receptors with AlphaFold2 (2022, methodology): “…the targets, each model’s similarity to the experimental structures was quantified using template modeling (TM) score ( Zhang and Skolnick, 2004 ; Zhang and Skolnick, 2005 ; Xu and Zhang, 2010 ), a metric ranging from 0 to 1, which indicates how well the two backbone (C α atoms) structures…”
Introducing “best single template” models as reference baseline for the Continuous Automated Model Evaluation (CAMEO) (2019, methodology): “For the “ bestSingleTemplate ” method, templates are discovered by structural superposition of the target reference structures with all PDB structures using TM-align.”
I-TASSER server: new development for protein structure and function predictions (2015, methodology): “The first I-TASSER model is searched against the PDB library by TM-align (31) to find the analogs that are structurally similar to the query proteins.”
ECOD: An Evolutionary Classification of Protein Domains (2014, background): “similarity and TMalign to reflect structure similarity [21,40].”
Learning protein sequence embeddings using information from structure (2019, methodology): “The largest improvement comes at the SCOP class level where TMalign achieves much lower average precision score for retrieving these weak structural matches.”
DeepSF: deep convolutional neural network for mapping protein sequences to folds (2017, methodology): “the models generated by top templates is selected for comparison. The quality of best predicted models from DeepSF and HHSearch is evaluated against the native structure in terms of TM-score and RMSD [28]. Here, we mainly evaluated template ranking and protein structure prediction on the 96 template”
Hallucinating symmetric protein assemblies (2022, methodology): “TM-scores (protomer | oligomer) are indicated in parentheses, and the PDB IDs are reported in Table S2.”