A unified framework identifies new links between plasma lipids and diseases from electronic medical records across large-scale cohorts

Veturi, Yogasudha; Lucas, Anastasia; Bradford, Yuki; Hui, Daniel; Dudek, Scott M.; Theusch, Elizabeth; Verma, Anurag; Miller, Jason E.; Kullo, Iftikhar J.; Hákonarson, Hákon; Sleiman, Patrick; Schaid, Daniel J.; Stein, C. Michael; Edwards, Digna R. Velez; Qi, Feng; Wei, Wei Qi; Medina, Marisa W.; Krauss, Ronald M.; Hoffmann, Thomas J.; Risch, Neil; Voight, Benjamin F.; Rader, Daniel J.; Ritchie, Marylyn D.

doi:10.1038/s41588-021-00879-y

Cited by 21 publications

(20 citation statements)

References 69 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To follow up on novel TWAS-prioritized genes, the PheWAS framework can be expanded to the GReX level in a similar analysis: GReX-level phenome-wide association study (GReX-PheWAS), similar to the PredixVU database. 56 – 59 Not only do these analyses replicate and detect new TWAS associations, they can also point to groups of phenotypes that show enrichments for trait associations for the gene of interest.…”

Section: Resultsmentioning

confidence: 99%

Best practices for multi-ancestry, meta-analytic transcriptome-wide association studies: Lessons from the Global Biobank Meta-analysis Initiative

Bhattacharya¹,

Hirbo²,

Zhou³

et al. 2022

Cell Genomics

View full text Add to dashboard Cite

Section: Resultsmentioning

confidence: 99%

Best practices for multi-ancestry, meta-analytic transcriptome-wide association studies: Lessons from the Global Biobank Meta-analysis Initiative

Bhattacharya¹,

Hirbo²,

Zhou³

et al. 2022

Cell Genomics

View full text Add to dashboard Cite

“…Additionally, for the most part, each function is able to be downloaded, modified if needed, and run as a standalone script in R. For example, in Veturi et al 2021, we modified hudson code to generate Figure 3 in which we rotated the plot 90 degrees so that we could display two Hudson plots side by side to compare gene expression based PheWAS (Xpress-PheWAS) and SNP-based PheWAS in two datasets 16 . The package was designed this way intentionally to increase accessibility to researchers who wish to make small tweaks to their figures, but do not have as much experience with R package development.…”

Section: Methodsmentioning

confidence: 99%

hudson: A User-Friendly R Package to Extend Manhattan Plots

Lucas

Verma

Ritchie

2022

Preprint

Self Cite

View full text Add to dashboard Cite

The Manhattan plot is one of the most widely used visualization techniques when plotting summary statistics from genome-wide or phenome-wide association studies. While there are a number of existing tools to create these plots, there is room for extending their utility to satisfy increasingly complex and comprehensive analyses as well as the need for comparisons between different sets of results or between discovery and replication datasets. The R package presented here, hudson, provides user-friendly plotting functions intended for use with genome, phenome, and exposome-wide association studies, but its flexible framework can be utilized for a wide variety genome-wide analyses. Further, we have extended these figures to allow for interactive elements to facilitate results exploration for the ever-increasing large-scale dimensionality of these data. hudson can be obtained from github.com/RitchieLab/hudson.

show abstract

“…The federated model helps construct a highly collaborative community that attracts more participating sites and enables rapid development and deployment of research projects across institutions. In addition, we also see the success of some flexible hybrid models, especially when collaborations are built across networks 56 . Consequently, flexible methods that can be efficiently implemented in different research networks are needed.…”

Section: Introductionmentioning

confidence: 91%

COMMUTE: communication-efficient transfer learning for multi-site risk prediction

Duan

2022

Preprint

View full text Add to dashboard Cite

Objectives: We propose a synthetic-data-based transfer learning approach (SynTL) to incorporate multi-center healthcare data for improving the risk prediction of a target population, accounting for challenges including heterogeneity, data sharing, and privacy constraints. Methods: SynTL combines locally trained risk prediction models from each source population with the target population to adjust for the data heterogeneity through a flexible distance-based transfer learning approach. Heterogeneity-adjusted synthetic data are then generated for source populations where individual-level data are not shareable. The synthetic data are then combined with the target and source data for joint training of the target model. We evaluate SynTL via extensive simulation studies and an application to multi-center data from the electronic Medical Records and Genomics (eMERGE) Network to predict extreme obesity. Results: Simulation studies show that SynTL outperforms methods without adjusting for the data heterogeneity and methods that are trained in a single population over a wide spectrum of settings. SynTL has low communication costs where each participating site only needs to share parameter estimates to the target, requiring only one round of communication. SynTL protects against negative transfer when some source populations are highly different from the target. Using eMERGE data, SynTL achieves an area under the receiver operating characteristic curve (AUC) around 0.79, which outperforms other benchmark methods (0.50 - 0.67). Conclusion: SynTL improves the risk prediction performance of the target population, and is robust to the level of heterogeneity between the target and source populations. It protects patient-level information and is highly communication efficient.

show abstract

A unified framework identifies new links between plasma lipids and diseases from electronic medical records across large-scale cohorts

Cited by 21 publications

References 69 publications

Best practices for multi-ancestry, meta-analytic transcriptome-wide association studies: Lessons from the Global Biobank Meta-analysis Initiative

Best practices for multi-ancestry, meta-analytic transcriptome-wide association studies: Lessons from the Global Biobank Meta-analysis Initiative

hudson: A User-Friendly R Package to Extend Manhattan Plots

COMMUTE: communication-efficient transfer learning for multi-site risk prediction

Contact Info

Product

Resources

About