pith. sign in

arxiv: 2604.08217 · v1 · submitted 2026-04-09 · 💻 cs.CY

Co-design for Trustworthy AI: An Interpretable and Explainable Tool for Type 2 Diabetes Prediction Using Genomic Polygenic Risk Scores

Ralf Beuthan , Megan Coffee , Heejin Kim , Na Yeon Kim , Pedro Kringen , Elisabeth Hildt , Haekyung Lee , Seunggeun Lee
show 8 more authors
Emilie Wiinblad Mathez Sira Maliphol Vadim Pak Yuna Park Stephan Sonnenberg Jesmin Jahan Tithi Magnus Westerlund Roberto V. Zicari
This is my paper

Pith reviewed 2026-05-10 17:58 UTC · model grok-4.3

classification 💻 cs.CY
keywords polygenic risk scorestype 2 diabetesSHAP explanationsexplainable AItrustworthy AIco-designgenomic predictionclinical AI tools
0
0 comments X

The pith

A visualization tool decomposes polygenic risk scores for type 2 diabetes into specific gene and SNP contributions using SHAP, developed through ethical co-design.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper presents XPRS, a tool designed to overcome the interpretability gap in polygenic risk scores for type 2 diabetes by breaking scores down to gene-level and single-nucleotide polymorphism contributions. Researchers applied co-design methods to embed considerations of ethics, law, human rights, and technical robustness from the outset. A sympathetic reader would care because clearer explanations of genetic risk could support more informed clinical decisions, earlier interventions, and greater trust in AI-assisted predictions. The work yields concrete lessons from a multidisciplinary team that can inform similar tools in other medical domains.

Core claim

The paper establishes that the XPRS tool, built on Shapley Additive Explanations, decomposes polygenic risk scores for type 2 diabetes into granular gene-level and SNP contribution scores, delivering individual risk insights. This is paired with a co-design process using Z-inspection and HUDERIA frameworks to evaluate and address legal, medical, ethical, and technical trustworthiness issues during development and potential use, producing a set of lessons learned that serve as guidance for future explainability frameworks in clinical PRS applications.

What carries the argument

The XPRS visualization tool that applies SHAP to decompose PRS into gene-level and SNP contribution scores, integrated with Z-inspection and HUDERIA co-design methods for assessing trustworthiness across ethical, legal, and technical dimensions.

If this is right

  • Clinicians gain granular views of which genetic factors drive an individual's type 2 diabetes risk, enabling more targeted screening and interventions.
  • Developers of similar PRS models receive a practical reference for building explainability into genomic tools.
  • Multidisciplinary teams obtain documented lessons on navigating ethical, legal, and robustness challenges when deploying AI in healthcare.
  • The approach supplies a template for creating trustworthy AI systems in other polygenic disease contexts.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Wider adoption of such decomposable PRS tools could help overcome hesitation around using genetic risk scores in routine medical care.
  • The same SHAP decomposition strategy might transfer to polygenic predictions for conditions like obesity or certain cancers.
  • Real-world testing in clinical workflows would be needed to confirm whether the added explanations actually change patient or physician behavior.

Load-bearing premise

That SHAP-based breakdowns of polygenic risk scores produce explanations that are biologically accurate and clinically useful rather than artifacts of the model or data correlations.

What would settle it

An independent study showing that the specific gene and SNP contributions highlighted by XPRS for high-risk patients do not align with established biological pathways or functional variants linked to type 2 diabetes.

Figures

Figures reproduced from arXiv: 2604.08217 by Elisabeth Hildt, Emilie Wiinblad Mathez, Haekyung Lee, Heejin Kim, Jesmin Jahan Tithi, Magnus Westerlund, Megan Coffee, Na Yeon Kim, Pedro Kringen, Ralf Beuthan, Roberto V. Zicari, Seunggeun Lee, Sira Maliphol, Stephan Sonnenberg, Vadim Pak, Yuna Park.

Figure 1
Figure 1. Figure 1: Z-Inspection® Co-design Process, adapted from [13]. The figure illustrates the iterative stages of the Z-Inspection® co-design methodology applied in this use case. Sociotechnical Scenarios Sociotechnical scenarios are essential in the Z-Inspection® process, linking the technical capabilities of AI systems with their societal impacts. They uncover potential ethical, technical, and legal issues while foster… view at source ↗
read the original abstract

The polygenic risk scores (PRS) have emerged as an important methodology for quantifying genetic predisposition to complex traits and clinical disease. Significant progress has been made in applying PRS to conditions such as obesity, cancer, and type 2 diabetes (T2DM). Studies have demonstrated that PRS can effectively identify individuals at high risk, thereby enabling early screening, personalized treatment, and targeted interventions for diseases with a genetic predisposition. One current limitation of PRS, however, is the lack of interpretability tools. To address this problem for T2DM, researchers at the Graduate School of Data Science at the Seoul National University introduced eXplainable PRS (XPRS). This visualization tool decomposes PRSs into gene-level and single-nucleotide polymorphism (SNP) contribution scores via Shapley Additive Explanations (SHAP), providing granular insights into the specific genetic factors driving an individual's risk profile. We used a co-design approach to assess XPRS trustworthiness by considering legal, medical, ethical, and technical robustness during early design and potential clinical use. For that, we used Z-inspection, an ethically aligned Trustworthy AI co-design methodology, and piloted the Council of Europe's Human Rights, Democracy, and the Rule of Law Impact Assessment for AI Systems (HUDERIA) (Council of Europe (CAI) 2025). The findings of this use-case comprise a comprehensive set of ethical, legal, and technical lessons learned. These insights, identified by a multidisciplinary team of experts (ethics, legal, human rights, computer science, and medical), serve as a framework for designers to navigate future challenges with this and other AI systems. The findings also provide a useful reference for researchers developing explainability frameworks for PRS in diverse clinical contexts.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The paper introduces eXplainable PRS (XPRS), a visualization tool that applies Shapley Additive Explanations (SHAP) to decompose polygenic risk scores (PRS) for Type 2 Diabetes (T2DM) into gene-level and SNP contribution scores. It describes a co-design process using Z-inspection and the HUDERIA framework to evaluate the tool's trustworthiness across legal, medical, ethical, and technical dimensions, and reports lessons learned from a multidisciplinary expert team.

Significance. If the SHAP decompositions can be shown to yield biologically grounded explanations rather than model artifacts, the work would offer a practical framework for interpretable genomic AI tools and a reusable co-design template for trustworthy AI in clinical contexts, addressing a recognized gap in PRS interpretability.

major comments (2)
  1. [Abstract] Abstract: the central claim that XPRS 'decomposes PRSs into gene-level and single-nucleotide polymorphism (SNP) contribution scores via SHAP, providing granular insights into the specific genetic factors driving an individual's risk profile' is presented without any performance metrics, validation against known T2DM loci or pathways, comparison to existing PRS tools, or empirical checks (e.g., eQTL overlap or perturbation tests) that would establish the attributions are biologically faithful rather than statistical artifacts or linkage-disequilibrium effects.
  2. [XPRS Tool and Co-design Assessment] Description of the XPRS tool and co-design process: the manuscript asserts that the Z-inspection and HUDERIA assessments identified relevant trustworthiness issues, yet supplies no concrete details on how SHAP outputs were evaluated for accuracy, stability, or clinical meaningfulness, leaving the reported ethical/legal/technical lessons dependent on an untested interpretability premise.
minor comments (1)
  1. [Abstract] The citation 'Council of Europe (CAI) 2025' appears to reference a future document; clarify the exact reference and publication status.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their detailed and constructive feedback, which highlights important considerations for strengthening the manuscript's claims. We appreciate the recognition of the work's potential contribution to interpretable genomic AI and trustworthy AI co-design. We address each major comment below and will revise the manuscript accordingly to clarify scope, temper claims, and add limitations without misrepresenting the current study.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the central claim that XPRS 'decomposes PRSs into gene-level and single-nucleotide polymorphism (SNP) contribution scores via SHAP, providing granular insights into the specific genetic factors driving an individual's risk profile' is presented without any performance metrics, validation against known T2DM loci or pathways, comparison to existing PRS tools, or empirical checks (e.g., eQTL overlap or perturbation tests) that would establish the attributions are biologically faithful rather than statistical artifacts or linkage-disequilibrium effects.

    Authors: We agree that the abstract phrasing risks overstating the tool's outputs as biologically faithful insights without supporting validation. The manuscript's core focus is the XPRS visualization tool and the multidisciplinary co-design process (via Z-inspection and HUDERIA) to derive ethical, legal, and technical lessons for trustworthy AI, rather than a technical validation study of SHAP attributions. We will revise the abstract to qualify the claim, emphasizing decomposition for visualization and user exploration instead of 'granular insights into specific genetic factors.' We will also add a limitations section explicitly noting the absence of performance metrics, locus validation, comparisons, or perturbation tests, and recommending these as directions for future work. This addresses the concern while preserving the paper's emphasis on the co-design framework. revision: yes

  2. Referee: [XPRS Tool and Co-design Assessment] Description of the XPRS tool and co-design process: the manuscript asserts that the Z-inspection and HUDERIA assessments identified relevant trustworthiness issues, yet supplies no concrete details on how SHAP outputs were evaluated for accuracy, stability, or clinical meaningfulness, leaving the reported ethical/legal/technical lessons dependent on an untested interpretability premise.

    Authors: The co-design evaluation applied Z-inspection and HUDERIA to assess the tool's overall trustworthiness dimensions (legal, medical, ethical, technical) in a hypothetical clinical context, drawing on expert input from multiple disciplines. It did not include quantitative technical validation of SHAP (e.g., accuracy or stability metrics), as that falls outside the paper's scope of reporting lessons from the process itself. We will revise the relevant sections to provide more concrete examples from the expert discussions, such as specific concerns raised about SHAP's potential for misleading attributions due to linkage disequilibrium and how these informed the derived lessons. We will also clarify that the lessons are process-oriented rather than dependent on proven biological fidelity of the attributions. revision: partial

Circularity Check

0 steps flagged

No circularity: descriptive development of XPRS tool and co-design assessment

full rationale

The manuscript presents the XPRS visualization tool that applies SHAP to decompose PRS into gene- and SNP-level contributions for T2DM risk, followed by a co-design trustworthiness evaluation using the external Z-inspection framework and HUDERIA impact assessment. No equations, fitted parameters, predictions, or derivation steps appear in the provided text. The central claims rest on the described implementation process and multidisciplinary lessons learned rather than any reduction of outputs to inputs by construction, self-citation chains, or renamed empirical patterns. The work is therefore self-contained with no load-bearing circular steps.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 1 invented entities

The work rests on domain assumptions about the validity of SHAP for genomic models and the completeness of the chosen ethical frameworks rather than new mathematical axioms or invented physical entities.

axioms (2)
  • domain assumption SHAP values provide meaningful and trustworthy explanations for PRS models in genomic data
    Invoked in the tool design without reported validation against known biology or clinical outcomes.
  • domain assumption Z-inspection and HUDERIA frameworks are adequate to assess all relevant trustworthiness aspects for this AI system
    The co-design methodology relies on these specific tools capturing legal, medical, ethical, and technical issues.
invented entities (1)
  • XPRS tool no independent evidence
    purpose: To decompose PRS into interpretable gene- and SNP-level contributions for T2DM
    New visualization tool introduced by the authors.

pith-pipeline@v0.9.0 · 5699 in / 1545 out tokens · 54876 ms · 2026-05-10T17:58:47.022382+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

62 extracted references · 62 canonical work pages

  1. [1]

    The National Assembly of Korea adopted the AI Basic Act in December 2024.6 With this passage, Korea became the second country with an AI-specific law that encompassed a comprehensive regulatory scope. Since the government revealed initial drafts of the enforcement decree and guidelines for this Act in September 2025, the contents of many provisions will s...

    work page 2024

  2. [2]

    7 See (NIA Korea,

    consolidates 19 different AI bills introduced to the National Assembly since May 2024, after the 22nd session of the National Assembly began. 7 See (NIA Korea,

    work page 2024

  3. [3]

    Discussion and validation of the identified points in the full WG1. Human Rights Law findings Positive contributions WG1 first stressed the importance of considering what potential positive contribution the XPRS AI system could have towards the achievement of human rights and not just any potential threats or human rights violations. Two human rights were...

    work page 1948

  4. [4]

    [32], prepared by the Ministry of Science and ICT (MIST) and adopted in December 2020, were designed to introduce general principles applicable across different fields. They consist of three principles (must be considered in the process of developing and using AI for humanity) and ten requirements (to follow those three principles, key requirements that m...

    work page 2020

  5. [5]

    trustworthy AI

    further complicates the transparency and availability of such data. Additionally, the availability of information generated by XPRS may also prove a liability that requires proper management. From these perspectives, accountability falls well outside the clinician-patient relationship. While WG1 explored some policies in Korea, the EU, and other jurisdict...

    work page 2023

  6. [6]

    Sun H, Saeedi P, Karuranga S, et al., IDF Diabetes Atlas: Global, regional and country-level diabetes prevalence estimates for 2021 and projections for

    work page 2021

  7. [7]

    2022;183:109119

    Diabetes Res Clin Pract. 2022;183:109119. doi:10.1016/j.diabres.2021.109119

    work page doi:10.1016/j.diabres.2021.109119 2022

  8. [8]

    Prediction of incident diabetes mellitus in middle-aged adults: the Framingham Offspring Study

    Wilson PWF, Meigs JB, Sullivan L, Fox CS, Nathan DM, D’Agostino RB. Prediction of incident diabetes mellitus in middle-aged adults: the Framingham Offspring Study. Arch Intern Med. 2007;167(10):1068-1074. doi:10.1001/archinte.167.10.1068

    work page doi:10.1001/archinte.167.10.1068 2007

  9. [9]

    Development and validation of QDiabetes-2018 risk prediction algorithm to estimate future risk of type 2 diabetes: cohort study

    Hippisley-Cox J, Coupland C. Development and validation of QDiabetes-2018 risk prediction algorithm to estimate future risk of type 2 diabetes: cohort study. BMJ. 2017;359:j5019. doi:10.1136/bmj.j5019

    work page doi:10.1136/bmj.j5019 2018

  10. [10]

    eClinicalMedicine

    Lee H, et al., Prediction model for type 2 diabetes mellitus and its association with mortality using machine learning in three independent cohorts from South Korea, Japan, and the UK: a model development and validation study. eClinicalMedicine. 2025;80:103069. doi:10.1016/j.eclinm.2025.103069

    work page doi:10.1016/j.eclinm.2025.103069 2025

  11. [11]

    Nat Genet

    Mahajan A, et al., Multi-ancestry genetic study of type 2 diabetes highlights the power of diverse populations for discovery and translation. Nat Genet. 2022;54(5):560-572. doi:10.1038/s41588-022-01058-3

    work page doi:10.1038/s41588-022-01058-3 2022

  12. [12]

    Development and validation of a trans-ancestry polygenic risk score for type 2 diabetes in diverse populations

    Ge T, Chen CY, Ni Y, Feng YA, Smoller JW. Development and validation of a trans-ancestry polygenic risk score for type 2 diabetes in diverse populations. Genome Med. 2022;14(1):70. doi:10.1186/s13073-022-01074-2

    work page doi:10.1186/s13073-022-01074-2 2022

  13. [13]

    Nat Genet

    Khera AV, et al., Genome-wide polygenic scores for common diseases identify individuals with risk equivalent to monogenic mutations. Nat Genet. 2018;50(9):1219-1224. doi:10.1038/s41588-018-0183-z

    work page doi:10.1038/s41588-018-0183-z 2018

  14. [14]

    An Improved Genome-Wide Polygenic Score Model for Predicting the Risk of Type 2 Diabetes

    Liu W, Zhuang Z, Wang W, Huang T, Liu Z. An Improved Genome-Wide Polygenic Score Model for Predicting the Risk of Type 2 Diabetes. Front Genet. 2021;12:632385. doi:10.3389/fgene.2021.632385

    work page doi:10.3389/fgene.2021.632385 2021

  15. [15]

    Mars N, et al., Polygenic and clinical risk scores and their impact on age at onset and prediction of cardiometabolic diseases and common cancers. Nat Med. 2020;26:549-557. doi:10.1038/s41591-020-0800-0

    work page doi:10.1038/s41591-020-0800-0 2020

  16. [16]

    Kim NY, et al., The clinical relevance of a polygenic risk score for type 2 diabetes mellitus in the Korean population. Sci Rep. 2024;14:5749. doi:10.1038/s41598-024-55313-0

    work page doi:10.1038/s41598-024-55313-0 2024

  17. [17]

    Influence and role of polygenic risk score in the development of 32 complex diseases

    Liu Y, Wang H, Gao T, Yan Y, Wang T, Zheng C, Zeng P. Influence and role of polygenic risk score in the development of 32 complex diseases. J Glob Health. 2025;15:04071. Available at: https://pmc.ncbi.nlm.nih.gov/articles/PMC11893022/

    work page 2025

  18. [18]

    15 years of GWAS discovery: Realizing the promise

    Abdellaoui A, Yengo L, Verweij KJH, Visscher PM. 15 years of GWAS discovery: Realizing the promise. Am J Hum Genet. 2023;110(2):179-194. doi:10.1016/j.ajhg.2022.12.011

    work page doi:10.1016/j.ajhg.2022.12.011 2023

  19. [19]

    IEEE Transactions on Technology and Society

    Zicari RV, et al., Z-Inspection: A Process to Assess Trustworthy AI. IEEE Transactions on Technology and Society. 2021;2(2):83-97. doi:10.1109/TTS.2021.3066209. 49

    work page doi:10.1109/tts.2021.3066209 2021

  20. [20]

    Clinical utility of polygenic risk scores: a critical 2023 appraisal

    Koch S, Schmidtke J, Krawczak M, Caliebe A. Clinical utility of polygenic risk scores: a critical 2023 appraisal. J Community Genet. 2023;14(5):471-487. doi:10.1007/s12687-023-00645-z

    work page doi:10.1007/s12687-023-00645-z 2023

  21. [21]

    XPRS: a tool for interpretable and explainable polygenic risk score

    Kim NY, Lee S. XPRS: a tool for interpretable and explainable polygenic risk score. Bioinformatics. 2025;41(4):btaf143. doi:10.1093/bioinformatics/btaf143

    work page doi:10.1093/bioinformatics/btaf143 2025

  22. [22]

    Nat Genet

    Steinthorsdottir V, Thorleifsson G, Reynisdottir I, et al., A variant in CDKAL1 influences insulin response and risk of type 2 diabetes. Nat Genet. 2007;39(6):770-775. doi:10.1038/ng2054

    work page doi:10.1038/ng2054 2007

  23. [23]

    arXiv:2206.09887 [cs.CY]

    Zicari RV, et al., How to Assess Trustworthy AI in Practice. arXiv:2206.09887 [cs.CY]

    work page arXiv

  24. [24]

    Engineering-Based Design Methodology for Embedding Ethics in Autonomous Robots

    Robertson LJ, Abbas R, Alici G, Munoz A, Michael K. Engineering-Based Design Methodology for Embedding Ethics in Autonomous Robots. Proc. IEEE. 2019;107(3):582-599. doi:10.1109/JPROC.2018.2889678

    work page doi:10.1109/jproc.2018.2889678 2019

  25. [25]

    Available at: https://ntrs.nasa.gov/api/citations/20210012614/downloads/

    work page arXiv

  26. [26]

    Regulation (EU) 2024/1689 laying down harmonised rules on artificial intelligence (AI Act)

    European Parliament and Council of the European Union. Regulation (EU) 2024/1689 laying down harmonised rules on artificial intelligence (AI Act). Official Journal of the European Union, L series, 12 July

    work page 2024

  27. [27]

    Available at: https://rm.coe.int/cai-2024-16rev2-methodology-for-the-risk-and-impact-assessment-of-arti/1680b2a09f

    work page 2024

  28. [28]

    Vilnius, 5.IX.2024

    Council of Europe Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law. Vilnius, 5.IX.2024. Available at: https://rm.coe.int/1680afae3c. 50

    work page 2024

  29. [30]

    Available at: https://arxiv.org/pdf/2404.14366

    work page arXiv

  30. [31]

    International Covenant on Civil and Political Rights (ICCPR)

    United Nations General Assembly. International Covenant on Civil and Political Rights (ICCPR). Adopted 16 December 1966, entered into force 23 March

    work page 1966

  31. [32]

    Human Dignity, International Protection

    Petersen N. Human Dignity, International Protection. Max Planck Encyclopedia of Public International Law. Oxford University Press. Available at: https://opil.ouplaw.com/display/10.1093/law:epil/9780199231690/law-9780199231690-e809

    work page doi:10.1093/law:epil/9780199231690/law-9780199231690-e809

  32. [33]

    Brief Bioinform

    Schwarzerova J, Hurta M, Barton V, et al., A perspective on genetic and polygenic risk scores: advances and limitations and overview of associated tools. Brief Bioinform. 2024;25(3):bbae240. doi:10.1093/bib/bbae240

    work page doi:10.1093/bib/bbae240 2024

  33. [34]

    Advancements and limitations in polygenic risk score methods for genomic prediction: a scoping review

    Jayasinghe D, Eshetie S, Beckmann K, Benyamin B, Lee SH. Advancements and limitations in polygenic risk score methods for genomic prediction: a scoping review. Hum Genet. 2024;143(12):1401-1431. doi:10.1007/s00439-024-02716-8

    work page doi:10.1007/s00439-024-02716-8 2024

  34. [35]

    Challenges and Opportunities for Developing More Generalizable Polygenic Risk Scores

    Wang Y, Tsuo K, Kanai M, Neale BM, Martin AR. Challenges and Opportunities for Developing More Generalizable Polygenic Risk Scores. Annu Rev Biomed Data Sci. 2022;5:293-320. doi:10.1146/annurev-biodatasci-111721-074830

    work page doi:10.1146/annurev-biodatasci-111721-074830 2022

  35. [36]

    Transl Vis Sci Technol

    Campbell JP, Mathenge C, Cherwek H, et al., Artificial Intelligence to Reduce Ocular Health Disparities. Transl Vis Sci Technol. 2021;10(3):19. doi:10.1167/tvst.10.3.19

    work page doi:10.1167/tvst.10.3.19 2021

  36. [37]

    Elhussein A, Baymuradov U, NYGC ALS Consortium, et al., A framework for sharing of clinical and genetic data for precision medicine applications. Nat Med. 2024;30:3578-3589. doi:10.1038/s41591-024-03239-5

    work page doi:10.1038/s41591-024-03239-5 2024

  37. [38]

    Utilization of genetic information for medicines development and equitable benefit sharing

    Matsuyama K, Kurihara C, Crawley FP, Kerpel-Fronius S. Utilization of genetic information for medicines development and equitable benefit sharing. Front Genet. 2023;14:1085864. doi:10.3389/fgene.2023.1085864

    work page doi:10.3389/fgene.2023.1085864 2023

  38. [39]

    Genetic data sharing and privacy

    Sorani MD, Yue JK, Sharma S, Manley GT, Ferguson AR; TRACK TBI Investigators. Genetic data sharing and privacy. Neuroinformatics. 2015;13(1):1-6. doi:10.1007/s12021-014-9248-z

    work page doi:10.1007/s12021-014-9248-z 2015

  39. [40]

    doi:10.1101/2023.02.21.23286110. 51

    work page doi:10.1101/2023.02.21.23286110 2023

  40. [41]

    Nat Commun

    Sun J, Wang Y, Folkersen L, et al., Translating polygenic risk scores for clinical use by estimating the confidence bounds of risk prediction. Nat Commun. 2021;12:5276. doi:10.1038/s41467-021-25014-7

    work page doi:10.1038/s41467-021-25014-7 2021

  41. [42]

    Nat Commun

    Misra A, Truong B, Urbut SM, et al., Instability of high polygenic risk classification and mitigation by integrative scoring. Nat Commun. 2025;16:1584. doi:10.1038/s41467-025-56945-0

    work page doi:10.1038/s41467-025-56945-0 2025

  42. [43]

    Genetic data are not always personal, disaggregating the identifiability and sensitivity of genetic data

    Rahnasto J. Genetic data are not always personal, disaggregating the identifiability and sensitivity of genetic data. J Law Biosci. 2023;10(2):lsad029. doi:10.1093/jlb/lsad029

    work page doi:10.1093/jlb/lsad029 2023

  43. [44]

    Int J Environ Res Public Health

    Amorim M, Silva S, Machado H, et al., Benefits and Risks of Sharing Genomic Data for Research. Int J Environ Res Public Health. 2022;19(14):8788. doi:10.3390/ijerph19148788

    work page doi:10.3390/ijerph19148788 2022

  44. [45]

    Cell Genomics

    Rehm HL, Page AJH, Smith L, et al., GA4GH: International policies and standards for data sharing across genomic research and healthcare. Cell Genomics. 2021;1(2):100029. doi:10.1016/j.xgen.2021.100029

    work page doi:10.1016/j.xgen.2021.100029 2021

  45. [46]

    Nat Genet

    Evangelou E, Warren HR, Mosen-Ansorena D, et al., Genetic analysis of over 1 million people identifies 535 new loci associated with blood pressure traits. Nat Genet. 2018;50:1412-1425. doi:10.1038/s41588-018-0205-x

    work page doi:10.1038/s41588-018-0205-x 2018

  46. [47]

    Kuchenbaecker KB, Hopper JL, Barnes DR, et al., Risks of Breast, Ovarian, and Contralateral Breast Cancer for BRCA1 and BRCA2 Mutation Carriers. JAMA. 2017;317(23):2402-2416. doi:10.1001/jama.2017.7112

    work page doi:10.1001/jama.2017.7112 2017

  47. [48]

    2023;30(5):558-586

    J Atheroscler Thromb. 2023;30(5):558-586. doi:10.5551/jat.CR005

    work page doi:10.5551/jat.cr005 2023

  48. [49]

    Genet Med

    Deignan JL, Astbury C, Cutting GR, et al., CFTR variant testing: a technical standard of ACMG. Genet Med. 2020;22(8):1288-1295. doi:10.1038/s41436-020-0822-5

    work page doi:10.1038/s41436-020-0822-5 2020

  49. [50]

    To trust or not to trust? An assessment of trust in AI-based systems

    Omrani N, Rivieccio G, Fiore U, Schiavone F, Agreda SG. To trust or not to trust? An assessment of trust in AI-based systems. Technol Forecast Soc Change. 2022;181:121763

    work page 2022

  50. [51]

    Understanding Physician Attitudes Toward AI in Clinical Decision-Making: Cross-Sectional Study

    Alhazmi F. Understanding Physician Attitudes Toward AI in Clinical Decision-Making: Cross-Sectional Study. JMIR Form Res. 2025;9:e79730

    work page 2025

  51. [52]

    JMIR Hum Factors

    Tamori H, Yamashina H, Mukai M, et al., Acceptance of the use of artificial intelligence in medicine among Japan’s doctors and the public. JMIR Hum Factors. 2022;9(1):e24680

    work page 2022

  52. [53]

    J Med Internet Res

    Oh S, Kim JH, Choi SW, et al., Physician confidence in artificial intelligence: an online mobile survey. J Med Internet Res. 2019;21(3):e12422

    work page 2019

  53. [54]

    The effects of explainability and causability on perception, trust, and acceptance: Implications for explainable AI

    Shin D. The effects of explainability and causability on perception, trust, and acceptance: Implications for explainable AI. Int J Hum Comput Stud. 2021;146:102551

    work page 2021

  54. [55]

    Towards trustworthy medical AI ecosystems

    Herzog C, Blank S, Stahl BC. Towards trustworthy medical AI ecosystems. AI Soc. 2025;40(4):2119-2139. 52

    work page 2025

  55. [56]

    Clin Genet

    Donohue KE, Gooch C, Katz A, et al., Pitfalls and challenges in genetic test interpretation. Clin Genet. 2021;99(5):638-649. doi:10.1111/cge.13917

    work page doi:10.1111/cge.13917 2021

  56. [57]

    Campbell EG, Clarridge BR, Gokhale M, et al., Data Withholding in Academic Genetics. JAMA. 2002;287(4):473-480. doi:10.1001/jama.287.4.473

    work page doi:10.1001/jama.287.4.473 2002

  57. [58]

    J Am Coll Cardiol

    Inouye M, Abraham G, Nelson CP, et al., Genomic Risk Prediction of Coronary Artery Disease in 480,000 Adults. J Am Coll Cardiol. 2018;72(16):1883-1893. doi:10.1016/j.jacc.2018.07.079

    work page doi:10.1016/j.jacc.2018.07.079 2018

  58. [59]

    Am J Hum Genet

    Antoniou AC, Spurdle AB, Sinilnikova OM, et al., Common breast cancer-predisposition alleles are associated with breast cancer risk in BRCA1 and BRCA2 mutation carriers. Am J Hum Genet. 2008;82(4):937-948. doi:10.1016/j.ajhg.2008.02.008

    work page doi:10.1016/j.ajhg.2008.02.008 2008

  59. [60]

    Deep learning and principal-agent problems of algorithmic governance: The new materialism perspective

    Kim ES. Deep learning and principal-agent problems of algorithmic governance: The new materialism perspective. Technol Soc. 2020;63:101378

    work page 2020

  60. [61]

    Nat Commun

    Fahed AC, Wang M, Homburger JR, et al., Polygenic background modifies penetrance of monogenic variants for tier 1 genomic conditions. Nat Commun. 2020;11:3635. doi:10.1038/s41467-020-17374-3

    work page doi:10.1038/s41467-020-17374-3 2020

  61. [62]

    Front Genet

    Fritzsche MC, Akyuz K, Cano Abadia M, et al., Ethical layering in AI-driven polygenic risk scores. Front Genet. 2023;14:1098439. doi:10.3389/fgene.2023.1098439. 53 APPENDIX A. SOCIO-TECHNICAL SCENARIOS TEMPLATE

    work page doi:10.3389/fgene.2023.1098439 2023

  62. [63]

    Zicari (Z-lead) …………………………………………………………………………………

    Ethics oversight and/or approval Has the AI system already undergone an ethical assessment or other approval? If not - why not? If so, was this internal/external, volunteer/regulated, and what was covered? Did they get a waiver? Was there a clearing, but it was very light or internal and not considered sufficient? 55 APPENDIX B LOG ……………………………………………………………...

    work page 2025