Skip to content
DraconiticResearch
Search Sources API
186,125 studies31 sourcesUpdated 47 min ago
Peer-reviewedCC BY 4.0Bioinformatics

GROMTools: scalable individual-level GReX imputation for mega-biobank-scale cohorts

  • Marios Anyfantakis1,2,3,4,5,6,7
  • Sanan Venkatesh1,2,3,4,5,6
  • Jamie J R Bennett1,2,3,4,5,6
  • Chenyu Liu1,2,3,4,5,6
  • Gabriel E Hoffman1,2,3,4,5,6
  • Panos Roussos1,2,3,4,5,6
  • Georgios Voloudakis1,2,3,4,5,6,8
8 affiliations
  1. 1Department of Psychiatry, Icahn School of Medicine at Mount Sinai , New York, NY,
  2. 2Center for Disease Neurogenomics, Icahn School of Medicine at Mount Sinai , New York, NY,
  3. 3Friedman Brain Institute, Icahn School of Medicine at Mount Sinai , New York, NY,
  4. 4Department of Genetics and Genomic Sciences, Icahn School of Medicine at Mount Sinai , New York, NY,
  5. 5Mental Illness Research, Education, and Clinical Center (VISN 2 South), James J. Peters VA Medical Center , Bronx, NY,
  6. 6Center for Precision Medicine and Translational Therapeutics, James J. Peters VA Medical Center , Bronx, NY,
  7. 7School of Medicine, University of Crete , Heraklion,
  8. 8Department of Artificial Intelligence and Human Health, Icahn School of Medicine at Mount Sinai , New York, NY,
Open at sourcePDFDOICite

Abstract

Motivation: The computational burden of individual-level genetically regulated gene expression (GReX) imputation has risen sharply with the growth of human mega-biobanks and the rapid expansion of transcriptomic imputation models across tissues and single-cell hierarchies. Existing tools were not designed for this setting and require complex, memory-intensive workflows that are poorly matched to shared and cloud-based compute environments, where runtime, memory, and I/O directly determine cost and throughput. GROMTools is an open-source C++ engine with an R interface that exploits sparse prediction weights, streams PLINK2 genotypes, and writes compact binary outputs for scalable individual-level GReX imputation.

Results: In UK Biobank (UKB), benchmarks on chromosome 1 across 50,000–450,000 individuals, 388,017 variants, and 11,724 gene-model pairs from 32 single-nucleus transcriptome imputation models, GROMTools produced near-identical predictions to PrediXcan and PLINK2, with minimum Pearson correlation >0.999 and maximum RMSE <0.001 across all of the imputed genes, while reducing CPU time by about 100-fold and peak memory by about 33-fold. These gains make routine biobank-scale individual-level GReX imputation practical and cost-efficient on standard CPU infrastructure.

Availability and implementation: GROMTools is freely available at https://github.com/voloudakislab/gromtools under the GPL v3 license. Documentation is available at https://voloudakislab.github.io/gromtools/. All scripts used for data quality control and benchmarking, together with log files and a snapshot of the GROMTools code version used for all of our analyses, are archived at DOI: https://doi.org/10.5281/zenodo.21926165.

Related

  1. GCTAg: scalable mixed-model analysis for biobank-scale agricultural cohorts
    Alexander S. Leonard, Qiongyu He, Natasha Watson +2
    arXivPreprintarXiv licenceFull textBioinformatics & computational biologyGenetics & genomicsarXiv
  2. QuImputer: an LD-informed QUBO/Ising formulation of genotype imputation with quantum-hardware execution
    M. Kim, T.-H. Lee
    bioRxivPreprintAll rights reservedFull textBioinformatics & computational biologybioRxiv
  3. A Unified Framework for Batch Correction and Imputation in Modern Mass Spectrometry Proteomics
    A. M. Anwar, S. Bayoumi, L. Lahti +1
    bioRxivPreprintCC BYFull textBioinformatics & computational biologybioRxiv
  4. Imputation-Based Harmonization Mitigates Site Effects Without Data Leakage in Machine Learning Studies
    N. Hillman, A. Chen, F. Hu +7
    bioRxivPreprintCC BYFull textNeurosciencebioRxiv
  5. Genotype imputation improves GWAS resolution and facilitates candidate gene identification in pearl millet
    K. Kambara, S. K. Gupta, D. Tsugama
    bioRxivPreprintCC BYFull textGenetics & genomicsbioRxiv

Cite

BibTeXRIS
BibTeX
@article{anyfantakis2026gromtools,
  title = {{GROMTools: scalable individual-level GReX imputation for mega-biobank-scale cohorts}},
  author = {Marios Anyfantakis and Sanan Venkatesh and Jamie J R Bennett and Chenyu Liu and Gabriel E Hoffman and Panos Roussos and Georgios Voloudakis},
  journal = {Bioinformatics},
  year = {2026},
  month = oct,
  doi = {10.1093/bioinformatics/btag719},
  url = {https://academic.oup.com/bioinformatics/article/doi/10.1093/bioinformatics/btag719/8859108},
  note = {CC BY 4.0},
}
RIS
TY  - JOUR
TI  - GROMTools: scalable individual-level GReX imputation for mega-biobank-scale cohorts
AU  - Marios Anyfantakis
AU  - Sanan Venkatesh
AU  - Jamie J R Bennett
AU  - Chenyu Liu
AU  - Gabriel E Hoffman
AU  - Panos Roussos
AU  - Georgios Voloudakis
PY  - 2026
DA  - 2026/10/01
JO  - Bioinformatics
DO  - 10.1093/bioinformatics/btag719
UR  - https://academic.oup.com/bioinformatics/article/doi/10.1093/bioinformatics/btag719/8859108
AB  - Motivation: The computational burden of individual-level genetically regulated gene expression (GReX) imputation has risen sharply with the growth of human mega-biobanks and the rapid expansion of transcriptomic imputation models across tissues and single-cell hierarchies. Existing tools were not designed for this setting and require complex, memory-intensive workflows that are poorly matched to shared and cloud-based compute environments, where runtime, memory, and I/O directly determine cost and throughput. GROMTools is an open-source C++ engine with an R interface that exploits sparse prediction weights, streams PLINK2 genotypes, and writes compact binary outputs for scalable individual-level GReX imputation.  Results: In UK Biobank (UKB), benchmarks on chromosome 1 across 50,000–450,000 individuals, 388,017 variants, and 11,724 gene-model pairs from 32 single-nucleus transcriptome imputation models, GROMTools produced near-identical predictions to PrediXcan and PLINK2, with minimum Pearson correlation >0.999 and maximum RMSE <0.001 across all of the imputed genes, while reducing CPU time by about 100-fold and peak memory by about 33-fold. These gains make routine biobank-scale individual-level GReX imputation practical and cost-efficient on standard CPU infrastructure.  Availability and implementation: GROMTools is freely available at https://github.com/voloudakislab/gromtools under the GPL v3 license. Documentation is available at https://voloudakislab.github.io/gromtools/. All scripts used for data quality control and benchmarking, together with log files and a snapshot of the GROMTools code version used for all of our analyses, are archived at DOI: https://doi.org/10.5281/zenodo.21926165.
SN  - 1367-4803
SN  - 1367-4811
ER  -