🏆 Finalist — NIH Data Sharing Index (“S-Index”) Challenge
Demo corpus. Scores are computed on a select set of biomedical paper/datasets and may be inaccurate for papers outside this corpus — DataRank relies on network effects that improve with scale. We aim to expand this into a fully open resource pending additional funding.

A community effort to optimize sequence-based deep learning models of gene regulation

Nature Biotechnology(2024)10.1038/s41587-024-02414-wSource: DataRank Database

A community effort to optimize sequence-based deep learning models of gene regulation is a dataset published in Nature Biotechnology (2024). On theSindex it has a DataRank of 0.593, placing it in the top 47.1% of the data-sharing corpus. It has been cited 29 times, with 20 citing works in its 1-hop citation network. Its calibrated FAIR score is 47/100.

Top 47%percentile
0.593DataRank
0.593Top 47%
Dataset Open Access29 citations · base score 3.2
Cite:
datarank_citation_only_1hop_v6· scope data_onlyMethodology

Abstract

A systematic evaluation of how model architectures and training strategies impact genomics model performance is needed. To address this gap, we held a DREAM Challenge where competitors trained models on a dataset of millions of random promoter DNA sequences and corresponding expression levels, experimentally determined in yeast. For a robust evaluation of the models, we designed a comprehensive suite of benchmarks encompassing various sequence types. All top-performing models used neural networks but diverged in architectures and training strategies. To dissect how architectural and training choices impact performance, we developed the Prix Fixe framework to divide models into modular building blocks. We tested all possible combinations for the top three models, further improving their performance. The DREAM Challenge models not only achieved state-of-the-art results on our comprehensive yeast dataset but also consistently surpassed existing benchmarks on Drosophila and human genomic datasets, demonstrating the progress that can be driven by gold-standard genomics datasets.

Data sources & pipeline
Pipeline:MetadataData-paper checkEnrichmentCitation networkScoring
Enrichment:Pending

FAIR Checklist

Context only (not used in score)
Findable (1/2)
  • Has DOI
Accessible (1/2)
  • Open Access
Interoperable (0/2)
    Reusable (1/3)
    • Dataset classification

    FAIR checklist signals are shown for context only and do not affect DataRank scoring.

    47FAIR score
    F Findable
    53
    A Accessible
    68
    I Interoperable
    25
    R Reusable
    42
    Top 56% by FAIRLLM-assessed✓ full text read

    Calibrated FAIR score — a parallel quality metric, independent of the DataRank citation score. See the full evaluation →

    DataRank Breakdown

    Base Score 81%Citation Network 19%

    Base Score Contribution

    0.483

    From this paper's citation signal

    Citation Network Contribution

    0.110

    From 10 citing papers with measurable signal

    Learn more about DataRank methodology →

    Top 5 citers driving the network score

    Ranked by citation count — the same ordering the engine uses when summing log1p(Cq) over citers.

    1. Long Short-Term Memory
      Neural Computation199794,983 citationsDataRank 1.7
    2. Fast gapped-read alignment with Bowtie 2
      Nature Methods201259,681 citationsDataRank 1.6
    3. ImageNet Large Scale Visual Recognition Challenge
      International Journal of Computer Vision201540,012 citationsDataRank 1.6
    4. Glove: Global Vectors for Word Representation
      Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP)201433,486 citationsDataRank 1.6
    5. Focal Loss for Dense Object Detection
      2017 IEEE International Conference on Computer Vision (ICCV)201725,041 citationsDataRank 1.5
    Why this DataRank?

    DataRank blends this paper's own citation count with the influence of the papers that cite it. Here, roughly 81% comes from its base citations and 19% from the citation network (10 citing papers contributed measurable signal).

    Base score B(p)
    log1p(citation_count) — grows sub-linearly, so a paper with 1,000 citations is not 10× a paper with 100.
    Network N(p)
    Σ over citers of log1p(Cq) ÷ max(outdegreeq, 1). Being cited by a highly-cited paper with few references counts most.
    Damping factor d = 0.85
    DataRank = (1−d)·B(p) + d·N(p) — the two cards above are each already multiplied by their share.
    Self-citations excluded
    Citers sharing any OpenAlex author ID with this paper are filtered out before the network sum.

    Citers are pulled from OpenAlex sorted by cited_by_count:descand capped per paper, so when the cap binds we keep the highest-signal references and the score is reproducible across reruns.

    Read the full methodology →

    Click a node to highlight its connections. Use scroll to zoom. Drag to pan.

    Node colors:CenterData PaperData + Open AccessNon-dataSelected & links| Node size = percentile rank

    Authors (111)

    Daria NoginaORCID,Dmitry PenzarORCID,Dohoon LeeORCID,Danyeong Lee,Nayeon Kim