Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2023 May 19;24(1):206.
doi: 10.1186/s12859-023-05338-5.

Accuracy of a machine learning method based on structural and locational information from AlphaFold2 for predicting the pathogenicity of TARDBP and FUS gene variants in ALS

Yuya Hatano  1 , Tomohiko Ishihara  2 , Osamu Onodera  1
Affiliations

Accuracy of a machine learning method based on structural and locational information from AlphaFold2 for predicting the pathogenicity of TARDBP and FUS gene variants in ALS

Yuya Hatano et al. BMC Bioinformatics. .

Abstract

Background: In the sporadic form of amyotrophic lateral sclerosis (ALS), the pathogenicity of rare variants in the causative genes characterizing the familial form remains largely unknown. To predict the pathogenicity of such variants, in silico analysis is commonly used. In some ALS causative genes, the pathogenic variants are concentrated in specific regions, and the resulting alterations in protein structure are thought to significantly affect pathogenicity. However, existing methods have not taken this issue into account. To address this, we have developed a technique termed MOVA (method for evaluating the pathogenicity of missense variants using AlphaFold2), which applies positional information for structural variants predicted by AlphaFold2. Here we examined the utility of MOVA for analysis of several causative genes of ALS.

Methods: We analyzed variants of 12 ALS-related genes (TARDBP, FUS, SETX, TBK1, OPTN, SOD1, VCP, SQSTM1, ANG, UBQLN2, DCTN1, and CCNF) and classified them as pathogenic or neutral. For each gene, the features of the variants, consisting of their positions in the 3D structure predicted by AlphaFold2, pLDDT score, and BLOSUM62 were trained into a random forest and evaluated by the stratified fivefold cross validation method. We compared how accurately MOVA predicted mutant pathogenicity with other in silico prediction methods and evaluated the prediction accuracy at TARDBP and FUS hotspots. We also examined which of the MOVA features had the greatest impact on pathogenicity discrimination.

Results: MOVA yielded useful results (AUC ≥ 0.70) for TARDBP, FUS, SOD1, VCP, and UBQLN2 of 12 ALS causative genes. In addition, when comparing the prediction accuracy with other in silico prediction methods, MOVA obtained the best results among those compared for TARDBP, VCP, UBQLN2, and CCNF. MOVA demonstrated superior predictive accuracy for the pathogenicity of mutations at hotspots of TARDBP and FUS. Moreover, higher accuracy was achieved by combining MOVA with REVEL or CADD. Among the features of MOVA, the x, y, and z coordinates performed the best and were highly correlated with MOVA.

Conclusions: MOVA is useful for predicting the virulence of rare variants in which they are concentrated at specific structural sites, and for use in combination with other prediction methods.

Keywords: AlphaFold2; Amyotrophic lateral sclerosis; MOVA; Missense variant; Prediction tool.

PubMed Disclaimer

Conflict of interest statement

There are no associations with companies or organizations that would constitute a conflict of interest requiring disclosure in relation to this study.

Figures

Fig. 1
Fig. 1
Work flowchart for MOVA. The x, y, z coordinates, and the plddt score for the amino acid residues at the substitution sites in the protein in the pdb file of the Alphafold2 database, and the ΔBLOSUM62 of the substituted amino acid residue, were used as parameters for random forest, XGBoost, or support vector machine (SVM) training (A). The sample group was randomly divided into five subsets as avoiding bias in objective variables. With one subset as the test cases and the rest as the training cases, we built the model. The predictions were calculated and validated using the test data. The models were iteratively built so that all five subsets were test cases. (B). The model was generated 30 times with all variants in the dataset as training data. The probability of each possible variant of the gene being pathogenic was predicted, and the average of the predictions was used as the MOVA value (C)
Fig. 2
Fig. 2
We used receiver operating characteristic (ROC) curve analysis to determine whether MOVA (red line), CADD (blue line), PolyPhen-2 (orange line), EVE (gray line), REVEL (black line), or AlphScore (green line) classified variants for TARDBP, FUS, SETX, TBK1, OPTN, SOD1, VCP, SQSTM1, ANG, UBQLN2, DCTN1, and CCNF as positive or negative. For MOVA, the stratified fivefold cross-validation was repeated 5 times, so the cvAUC function of the cvAUC package was used to draw the average of the ROC curves for 25 times
Fig. 3
Fig. 3
All variants of TARDBP and FUS in the dataset were divided into positive variants (red) associated with ALS and negative variants (light blue) recognized in the general population. For Polyphen-2, pph2_prob (classifier probability of the variation being damaging), and MOVA, the 5F-MV value was plotted on the y-axis as the predicted value and the residue number was plotted on the x-axis. Both pph2_prob and 5F-MV take values between 0 and 1, with 1 having the highest probability of being pathogenic
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

References

    1. Adzhubei IA, Schmidt S, Peshkin L, Ramensky VE, Gerasimova A, Bork P, Kondrashov AS, Sunyaev SR. A method and server for predicting damaging missense mutations. Nat Methods. 2010;7:248–249. doi: 10.1038/nmeth0410-248. - DOI - PMC - PubMed
    1. Akobeng AK. Understanding diagnostic tests 3: receiver operating characteristic curves. Acta Paediatr. 2007;96:644–647. doi: 10.1111/j.1651-2227.2006.00178.x. - DOI - PubMed
    1. Bolognesi B, Faure AJ, Seuma M, Schmiedel JM, Tartaglia GG, Lehner B. The mutational landscape of a prion-like domain. Nat Commun. 2019;10:4162. doi: 10.1038/s41467-019-12101-z. - DOI - PMC - PubMed
    1. Breiman L. Random forests. Mach Learn. 2001;45:5–32. doi: 10.1023/A:1010933404324. - DOI
    1. Cady J, Allred P, Bali T, Pestronk A, Goate A, Miller TM, Mitra RD, Ravits J, Harms MB, Baloh RH. Amyotrophic lateral sclerosis onset is influenced by the burden of rare variants in known amyotrophic lateral sclerosis genes. Ann Neurol. 2015;77:100–113. doi: 10.1002/ana.24306. - DOI - PMC - PubMed