Yao M, Miller GW, Vardarajan BN, Baccarelli AA, Guo Z, Liu Z. Deciphering proteins in Alzheimer's disease: A new Mendelian randomization method integrated with AlphaFold3 for 3D structure prediction. Cell Genom. 2024 Dec 11;4(12):100700. Epub 2024 Dec 4 PubMed.
Recommends
Please login to recommend the paper.
Comments
Emory University
Yao et al. propose an interesting Mendelian randomization (MR) method that incorporates a robust, instrumental-variable selection procedure and downstream integration with AlphaFold to predict the effects of missense variants on 3D structure. The authors’ algorithm addresses a number of challenges using MR with proteomics data. An advantage of their approach is the robust way protein quantitative trait loci (pQTLs) are selected as instrumental variables in MR. One issue to consider for this selection is genetic variation, which could lead to an apparent change in exposure/protein level, simply due to loss, or gain, of binding affinity of the reagent to the protein; this is less of a problem with direct methods, such as mass spectrometry. The authors do not comment on this issue in the article, but it would be interesting to see how their algorithm handles such instances.
A nice feature of the algorithm is incorporation of three-dimensional structure prediction with AlphaFold to see how missense variants may ultimately affect protein structure. Incorporation of co-localization analysis into the package, as the authors plan to do, will provide even more utility. The MR-SPI method represents an advance for proteomics-based MR. Congratulations to the authors.
View all comments by Erik JohnsonUniversity of Eastern Finland
Yao et al. have developed and tested a novel Mendelian randomization method that is integrated to the three-dimensional protein structure prediction tool. Specifically, the authors first established an MR method called MR-SPI that selects valid protein quantitative loci (pQTL). It is based on the principal of Leo Tolstoy’s dictum “all happy families are alike; each unhappy family is unhappy in its own way” to identify causal protein biomarkers for health outcomes. This MR framework selects valid pQTLs as instrumental variables to be subsequently used also for the three-dimensional structural protein analysis (AlphaFold3). After testing the performance of MR-SPI using simulated data alongside other established MR methods, the authors applied this all-in-one pipeline to discover potential causal plasma protein biomarkers associated with Alzheimer’s disease (AD). For that purpose, plasma-based proteomics data and ~23 million imputed autosomal variants across ~1,500 proteins from the U.K. Biobank were retrieved from a cohort comprising ~54,000 participants. After filtering steps to select independent and strongly associated pQTL instrument variables, the authors used summary statistics for clinically diagnosed AD and AD by proxy from a meta-analysis of GWAS’s originating from the Jansen et al. study (2019). As a result, MR-SPI identified seven plasma proteins (CD33, CD55, EPHA1, PILRA, PILRB, RET, and TREM2) that either were positively or negatively associated with the risk of AD. Furthermore, the following AlphaFold3 and gene ontology enrichment analyses revealed specific local protein changes as well as common biological readouts linked to identified plasma targets.
Collectively, the stepwise process of how the MPI-SPI framework selects, and validates by voting, the most relevant instrument variables to the final estimation of causal effects is extremely insightful. This provides an alternative strategy to identify potential AD-associated biomarkers.
Given the rapidly accumulating genetic and omics data in different databases and biobanks world-wide, it is of utmost importance to leverage these kinds of integrated MR pipelines to pinpoint disease-related proteins to be further applied for biomarker development. Although there are also some limitations noted by the authors, the MR-SPI pipeline indeed holds great potential not only for detecting causal protein biomarkers associated with diseases, but also for the characterization of underlying disease mechanisms.
References:
Jansen IE, Savage JE, Watanabe K, Bryois J, Williams DM, Steinberg S, Sealock J, Karlsson IK, Hägg S, Athanasiu L, Voyle N, Proitsi P, Witoelar A, Stringer S, Aarsland D, Almdahl IS, Andersen F, Bergh S, Bettella F, Bjornsson S, Brækhus A, Bråthen G, de Leeuw C, Desikan RS, Djurovic S, Dumitrescu L, Fladby T, Hohman TJ, Jonsson PV, Kiddle SJ, Rongve A, Saltvedt I, Sando SB, Selbæk G, Shoai M, Skene NG, Snaedal J, Stordal E, Ulstein ID, Wang Y, White LR, Hardy J, Hjerling-Leffler J, Sullivan PF, van der Flier WM, Dobson R, Davis LK, Stefansson H, Stefansson K, Pedersen NL, Ripke S, Andreassen OA, Posthuma D. Genome-wide meta-analysis identifies new loci and functional pathways influencing Alzheimer's disease risk. Nat Genet. 2019 Mar;51(3):404-413. Epub 2019 Jan 7 PubMed.
View all comments by Mikko HiltunenMake a Comment
To make a comment you must login or register.