Community Polygenic Lipid Load Feedback Dashboard for Public Cardiovascular Risk Awareness – Design and Rationale Using Synthetic Data

ABSTRACT

Background
The genetic architecture of plasma lipids and lipoproteins spans rare, highly penetrant variants causing monogenic dyslipidemias and a broad background of common variants contributing to polygenic risk. The work of Hegele and colleagues has characterized this spectrum, showing how monogenic disorders, common variants identified by genome-wide association studies, and combined lipid disturbances contribute to cardiovascular risk. Polygenic risk scores (PRS) for coronary artery disease (CAD) and lipid traits now achieve measurable predictive utility and can refine statin eligibility and lifetime risk estimation. In parallel, Achilles tendon xanthomas and tendon thickening remain specific but underused phenotypes for familial hypercholesterolemia (FH) diagnosis and risk stratification, amenable to ultrasound and MRI quantification. Risk communication literature indicates that visual and interactive tools can improve understanding, risk perception, and motivation for cardiovascular prevention without necessarily increasing anxiety.

Objective
To design and technically describe a synthetic-data prototype of a “Polygenic Lipid Load Feedback Dashboard” that (1) computes a simplified polygenic lipid risk score (PLRS) and related traits for a synthetic community cohort, (2) exports JSON suitable for interactive web visualization, and (3) renders three pastel, high-interactivity charts in Svelte/Chart.js, targeting public awareness and triage-oriented exploration rather than formal clinical risk prediction.

Methods
We generated N=200 synthetic individuals with age, body mass index (BMI), PLRS, percentile rank within the population, self-reported family history of premature cardiovascular disease, Achilles tendon thickness, and postal code. Distributions were chosen to reflect realistic but non-identifiable patterns inspired by published ranges for lipid traits and tendon thickness in FH, without replicating any real dataset. Data were synthesized in Python using NumPy, pandas, and Faker, and exported as (a) a cohort_summary.json file with PLRS distribution statistics, (b) a geo_summary.json file summarizing average PLRS by postal region, and (c) per-participant JSON files (IND_100000…IND_100199). A single-phase workflow in Google Colab writes these JSON files into Google Drive for subsequent ingestion into a SvelteKit frontend. The frontend renders: (1) a histogram of PLRS with an interactive “awareness threshold” slider and a vertical line marking an individual’s PLRS; (2) a horizontal bar “geo-heat” chart of average PLRS by postal code with a search box and threshold highlighting; and (3) a triage scatter plot of PLRS versus Achilles tendon thickness, with point size and outline indicating family history.

Results

The synthetic cohort exhibited a unimodal PLRS distribution approximating a standard normal with a wide percentile spread, enabling separation of low-, intermediate-, and high-polygenic-load groups. Achilles tendon thickness and PLRS were generated with partial positive correlation to enable visually meaningful triage plots, consistent with clinical data linking tendon thickness, LDL burden, and cardiovascular risk in FH. The histogram dashboard allowed users to adjust a percentile-based awareness threshold (e.g., top 10% polygenic risk) and observe dynamic recolouring of high-risk distribution tails, while a vertical line labeled “You” marked the individual PLRS value. The postal-code chart ranked regions by average PLRS and visually emphasized higher-risk areas via pastel gradations and thicker borders above a chosen percentile threshold. The triage scatter plot allowed users to filter by minimum PLRS percentile and optionally restrict to those with a positive family history, yielding an immediate visual impression of individuals combining elevated polygenic burden and tendon phenotype signals. All charts operated purely on synthetic JSON without any server-side computation, making them deployable to static hosting environments or Firebase.

Discussion
This prototype connects several strands of current evidence: (1) the complex genetic architecture of lipid traits described by Hegele and colleagues; (2) the emerging clinical utility of PRSs for CAD and lipids; (3) the diagnostic and prognostic relevance of Achilles tendon xanthomas and tendon thickness in FH; and (4) the growing evidence base supporting interactive visual risk communication and map-based public health dashboards. The dashboard is deliberately positioned as a public awareness and health informatics concept rather than a validated clinical tool; it demonstrates how polygenic information and tendon features might be integrated in a user-facing interface to support risk perception and triage thinking. Integration with real cohorts, ethical governance for genetic risk disclosure, and rigorous evaluation of behavioural impact remain future steps.

Conclusion
A community-oriented Polygenic Lipid Load Feedback Dashboard, even when prototyped using synthetic data, provides a technically coherent and conceptually grounded framework for combining PLRS, Achilles tendon information, and region-level aggregation into interactive charts. The system reflects contemporary genetic and imaging insights into dyslipidemia and FH while drawing on evidence that visual and dynamic risk communication tools can support awareness and behavioural intention. Future work should connect such dashboards to real biobank or clinic data under appropriate governance, assess user comprehension and behavioural outcomes, and explore how polygenic risk and tendon imaging outputs can be ethically and effectively communicated outside specialist clinics.

1. INTRODUCTION

1.1 Genetic architecture of dyslipidemia and Hegele’s contributions

Plasma lipids and lipoproteins, including LDL cholesterol, HDL cholesterol, and triglycerides, are influenced by both rare, large-effect variants and numerous common variants of small effect. Hegele’s work has documented how monogenic dyslipidemias, combined lipid disturbances, and GWAS-discovered loci together shape lipid phenotypes and cardiovascular risk. Up to 25 monogenic dyslipidemias involving at least 23 genes have been described, with implications for targeted therapies and cascade screening. At the same time, polygenic contributions can produce FH-like risk even in the absence of classic monogenic variants.

1.2 Polygenic risk scores for cardiovascular disease

Polygenic risk scores (PRSs) aggregate the effects of hundreds to millions of common variants into a single metric representing genetic liability for disease. Recent scientific statements and large cohort studies report that PRSs for CAD can meaningfully stratify risk, reclassify individuals relative to clinical scores, and help identify younger individuals at high lifetime risk who might benefit from earlier statin therapy. For example, a multi-ancestry CAD PRS (GPSMult) improved risk prediction across ancestries when integrated with conventional factors. A large study of over 330,000 individuals showed that a CAD PRS contributed substantially to myocardial infarction risk, especially in younger adults, and could refine statin initiation decisions. These developments suggest that polygenic risk may become increasingly relevant outside specialist lipid clinics.

1.3 Achilles tendon xanthomas and tendon thickness

Achilles tendon xanthomas, caused by cholesterol deposition within the tendon matrix, are highly specific for FH and associated with increased cardiovascular risk. Clinical palpation can miss subtle tendon changes; ultrasound and MRI improve detection and allow quantitative assessment of tendon thickness and echotexture. Recent reviews emphasize that Achilles tendon thickness measured by ultrasonography can aid FH diagnosis and risk assessment and note that some national criteria, such as those in Japan, explicitly incorporate tendon thickness. Observational work links greater Achilles tendon thickness with age, LDL cholesterol levels, and lipoprotein(a), reinforcing its role as a marker of cumulative lipid exposure and CVD risk.

1.4 Risk communication and interactive visual tools

Cardiovascular risk communication research indicates that the way risk is framed and visualized can materially influence patient understanding, perceived risk, and behavioural intention. Visual aids, including charts, icon arrays, and dynamic graphics, have been shown to improve comprehension without necessarily increasing anxiety or fatalism when designed well. Public health agencies have also adopted interactive dashboards and atlases—such as cardiovascular risk factor maps for Canadian regions—to communicate risk patterns and inequities at the population level.

1.5 Rationale for a polygenic lipid load feedback dashboard

Despite advances in lipid genetics and PRS methodology, most public-facing tools still focus on traditional risk factors (blood pressure, diabetes, smoking) and global risk calculations. There is little evidence that community-oriented tools explicitly represent polygenic lipid burden, tendon imaging markers, and regional patterns in a coherent interface. Existing work from Hegele’s group and colleagues demonstrates the complexity of genetic dyslipidemia and the value of tendon phenotyping in FH, while risk communication research shows that interactive visualization can support understanding and motivation. This project therefore focuses on the design and technical implementation of a synthetic-data dashboard that integrates a simplified PLRS with tendon thickness and regional information, aiming at public awareness and triage thinking rather than individualized clinical decision-making.

2. METHODS

2.1 Overall system design

The system comprises three layers:

1. Synthetic data generation in Python (Google Colab)
2. Export of JSON files to Google Drive for static hosting or Firebase storage
3. A SvelteKit + Chart.js frontend rendering three interactive charts powered entirely by client-side JSON ingestion

No real or identifiable patient data are used; all records are synthetic and designed for demonstration and methodology development.

2.2 Synthetic cohort generation

The Colab notebook mounts Google Drive and creates an output directory:

Code snippet 1: Google Drive mounting and directory preparation

from google.colab import drive
drive.mount('/content/drive')
import os, json, numpy as np, pandas as pd
from faker import Faker
BASE_DIR = "/content/drive/MyDrive/drb/Projects/Polygenic_load"
os.makedirs(BASE_DIR, exist_ok=True)
os.makedirs(f"{BASE_DIR}/participants", exist_ok=True)
fake = Faker()

A sample size of N=200 individuals was chosen for responsive front-end performance and visually interpretable plots. For each individual, the notebook simulates:
· Polygenic lipid risk score (PLRS) ~ N(0, 1)
· Percentile rank within the cohort, computed as empirical rank
· Age (18–74 years), BMI (approximately normal around 27.5 kg/m²)
· Family history of premature CVD (binary, probability 0.25)
· Achilles tendon thickness (approximately normal around 5.5 mm, modestly correlated with PLRS)
· Postal code prefix (Canada-like pseudo codes using Faker)

Code snippet 2: PLRS and percentile computation

N = 200
plrs = np.random.normal(loc=0.0, scale=1.0, size=N)
percentiles = pd.Series(plrs).rank(pct=True).values
ages = np.random.randint(18, 75, N)
bmi = np.random.normal(27.5, 3.5, N)
family_history = np.random.binomial(1, 0.25, N)
tendon_thickness = np.random.normal(5.5, 0.8, N)
postal_codes = [fake.postcode()[:3] for _ in range(N)]
df = pd.DataFrame({
"participant_id": [f"IND_{100000+i}" for i in range(N)],
"postal_code": postal_codes,
"plrs": plrs,
"percentile": percentiles,
"age": ages,
"bmi": bmi,
"family_history": family_history,
"tendon_thickness_mm": tendon_thickness
})

Percentile ranking mirrors approaches used in PRS distribution plots, where an individual’s score is contextualized within the population distribution. The tendon thickness distribution is loosely inspired by ranges reported in FH imaging studies but remains synthetic.

2.3 JSON schema and export

Each individual is written to a separate JSON file, creating a simple, file-based API for Svelte:

Code snippet 3: Per-participant JSON export

for _, row in df.iterrows():
    person = {
        "participant_id": row.participant_id,
        "plrs": float(row.plrs),
        "percentile": float(row.percentile),
        "age": int(row.age),
        "bmi": float(row.bmi),
        "family_history": int(row.family_history),
        "tendon_thickness_mm": float(row.tendon_thickness_mm),
        "postal_code": row.postal_code
    }
    with open(f"{BASE_DIR}/participants/{row.participant_id}.json", "w") as f:
        json.dump(person, f, indent=2)

In addition, cohort_summary.json stores basic statistics and arrays of PLRS values and percentiles for histogram visualization. A geo_summary.json file stores mean PLRS by postal region for the geo-heat chart, computed via groupby in pandas.

2.4 Frontend implementation (SvelteKit + Chart.js)

The frontend uses Svelte components located under src/lib/components/charts/Polygenic_load_feedback. Each component fetches the relevant JSON files from static/data/json/Polygenic_load_feedback (or corresponding Firebase bucket) at runtime, constructs the appropriate datasets, and configures Chart.js with pastel colour palettes, hover effects, click events, and sliders that dynamically modify dataset properties (e.g., bar colours or filtering criteria). Interactions are implemented entirely client-side using Svelte reactivity and Chart.js update mechanisms, without any live machine learning inference.

2.5 Ethics

Because all data are synthetically generated and non-identifiable, and no human subjects are involved, this prototype, in its current form, does not require research ethics approval. If adapted to real patient data in the future, ethics review, consent processes, and governance for genetic and imaging information would be essential.

3. RESULTS

Polygenic lipid risk in the community

3.1 Cohort distributions

The synthetic cohort exhibits:
· PLRS values spanning a range typical of standardized scores, with approximately symmetric distribution and some tail extension to represent high-polygenic-risk individuals.
· Percentiles uniformly distributed by design between 0 and 1, allowing consistent mapping to risk bands.
· Achilles tendon thickness values clustering around mid-single-digit millimetres with mild spread, consistent with values reported where increased thickness indicates FH in some populations.
· Family history positive in roughly one quarter of individuals, providing a meaningful subset for triage-focused visualization.

These structural properties make the cohort suitable for demonstrating the dashboard’s intended behaviours, even though it does not model real-world correlations with clinical outcomes.

3.2 PLRS cohort awareness chart

The first chart—PlrsCohortAwareness—renders a histogram of PLRS values with:
· 20 bins spanning the observed PLRS range
· Pastel blue bars for the bulk of the distribution, turning pastel red with darker borders once the interactive awareness threshold slider is set to a specific percentile (e.g., 90th percentile)
· A vertical line annotated “You” marking a chosen participant’s PLRS value, implemented via a custom Chart.js plugin that draws a line at x = PLRS (in chart coordinates) after datasets are rendered

As the user moves the awareness threshold slider, bars representing PLRS values above the corresponding percentile dynamically recolour, visually highlighting the “tail” of the distribution. This design echoes published work on visual risk communication, in which thresholds and green–yellow–red gradations are used to frame risk categories and help users locate themselves within a distribution. Clicking on a bar yields text feedback about the approximate PLRS range and number of individuals in that bin.

Regions and average polygenic lipid load

3.3 Regional polygenic load heat bar

The second chart, PlrsGeoHeatBar, mimics a choropleth by aggregating participants by postal-code region and rendering horizontal bars:
· Each bar’s length encodes mean PLRS for that region
· Each bar’s fill colour transitions from pastel teal (low mean PLRS) to pastel red (high mean PLRS), with darker outlines for higher values
· A risk-threshold slider dynamically adjusts which regions are visually emphasized via thicker borders, based on the percentile of the mean PLRS distribution
· A search box filters regions by partial postal-code strings, enabling quick lookup of specific areas

This representation is conceptually aligned with public health dashboards that display regional variation in cardiovascular risk factors or disease burden using map-based or bar-based visualizations. Although our implementation uses synthetic postal codes and summary scores, the structure illustrates how polygenic lipid load could be overlaid on geographical units in a health-informatics context.

Triage view: polygenic load and Achilles tendon thickness

3.4 Triage scatter: PLRS versus Achilles tendon thickness

The third chart, PlrsTriageScatter, plots individual points with:
· X-axis: PLRS
· Y-axis: Achilles tendon thickness
· Colour: polygenic risk band (e.g., pastel teal for low, pastel yellow/green for middle, pastel orange/red for high)
· Point radius and border width: larger and darker for individuals with a positive family history of premature CVD

User controls include:
· A slider adjusting the minimum percentile of PLRS displayed, effectively “zooming in” on individuals with higher genetic risk
· A checkbox restricting the display to those with family history, highlighting potential candidates for further assessment

This triage visualization is deliberately aligned with literature describing Achilles tendon thickness and xanthomas as markers of cumulative LDL exposure and cardiovascular risk in FH, and with studies indicating that PRS can identify individuals whose risk may resemble FH even without extreme LDL levels. By allowing dynamic filtering on these axes, the chart illustrates how a dashboard might conceptually support clinicians or public health practitioners in exploring combinations of polygenic burden, tendon phenotype, and family history in a cohort view.

4. DISCUSSION

4.1 Positioning within Hegele’s dyslipidemia framework

Hegele’s reviews emphasize that dyslipidemia traits are shaped by both monogenic disorders and a diffuse polygenic background, and that combined lipid disturbances carry distinct clinical implications. By operationalizing a simplified PLRS and tendon phenotype within a public-facing dashboard, this project translates those conceptual frameworks into a concrete, albeit synthetic, interface. While the prototype does not model specific genes or variant classes, it respects the notion that some individuals will carry higher polygenic load and potentially exhibit tendon changes even in community samples.

4.2 Polygenic risk scores and public-facing tools

Most current PRS work focuses on clinical integration—how to add PRS to existing risk scores, determine cutoffs for “high genetic risk,” and guide therapy decisions such as statin initiation or more intensive lifestyle interventions. Public-facing tools that display PRS directly to individuals are less well studied, particularly in the lipid domain. The dashboard here is explicitly not a clinical decision tool but an informatics experiment: it explores how polygenic information might be framed as a relative position in a population distribution and as part of a broader phenotype including tendon thickness and family history. Future empirical studies would be needed to evaluate whether such displays enhance understanding or motivation for prevention, and to manage potential harms (e.g., fatalism, anxiety, misinterpretation).

4.3 Achilles tendon imaging in a dashboard context

Recent work reinforces that Achilles tendon thickness and softness measured via ultrasound and advanced imaging can improve FH diagnosis and may track response to LDL-lowering therapy. Some countries already embed tendon thickness into FH diagnostic criteria. Incorporating a tendon phenotype axis into a triage scatter plot is therefore consistent with current scientific directions. While our prototype uses synthetic tendon thickness values, the structure of the chart—linking PLRS and tendon phenotype with family history flags—illustrates how a real dashboard could help flag high-priority individuals for specialist referral or imaging in a resource-constrained system.

4.4 Risk communication and behavioural intention

Systematic reviews and controlled studies on cardiovascular risk communication show that tailored and visual tools can improve risk perception and sometimes behavioural intentions, especially when they present comparative risk and trajectory information. Dashboards and maps used by public health agencies demonstrate that interactive visualizations can effectively communicate region-level risk and inequities. The dashboard in this project adopts similar principles: percentile positioning, clear thresholds, colour-banded risk strata, and region-level aggregation. However, rigorous evaluation—e.g., randomized trials comparing dashboard exposure versus standard information—would be necessary before deploying any similar tool in real-world settings that involve genetic risk disclosure.

4.5 Limitations

Key limitations include:

1. Synthetic data: All results are purely structural demonstrations; no clinical validity or generalizability can be inferred.
2. No outcome modelling: The dashboard does not link PLRS or tendon thickness to actual cardiovascular events or therapy response; it is not a risk calculator.
3. No usability or behavioural testing: The project does not include user studies, and therefore any benefit in awareness or behaviour remains speculative.
4. Governance and ethics: Real PRS and tendon imaging data require careful consideration of consent, counselling, data protection, and equity; these aspects are beyond the scope of the prototype.

5. CONCLUSION

This project presents a technically detailed, synthetic-data implementation of a Polygenic Lipid Load Feedback Dashboard inspired by contemporary dyslipidemia genetics, PRS research, Achilles tendon imaging, and risk communication science. It demonstrates how:
· A simplified PLRS distribution can be displayed with adjustable thresholds and individual positioning
· Regional aggregation of polygenic burden can be visualized in pastel, interactive charts
· A triage scatter plot can combine PLRS, tendon thickness, and family history in a way that echoes clinical reasoning about FH and related disorders

The system is not a clinical tool but a health-informatics prototype illustrating how these domains might converge in a public-facing interface. Future work could attach the same architecture to real cohorts under proper governance, integrate monogenic variant information (e.g., LipidSeq-based findings), and evaluate the impact of such dashboards on patient understanding, risk perception, and preventive behaviours.

FIGURE LEGENDS

Figure 1. Polygenic lipid risk distribution and individual positioning.
Histogram of synthetic polygenic lipid risk scores (PLRS) for N=200 individuals, with 20 bins spanning the observed range. Bars are rendered in a pastel blue palette, transitioning to pastel red with darker borders for bins exceeding an interactive percentile-based awareness threshold (e.g., top 10% PLRS). A custom vertical line marked “You” overlays the binning to indicate the PLRS of a selected individual. Users can adjust the awareness threshold via a slider, dynamically recolouring the high-risk tail and clicking on bins to reveal the approximate PLRS range and number of individuals in that bin. This figure illustrates population positioning of polygenic lipid load.

Figure 2. Regional average polygenic lipid load by postal-code region.
Horizontal bar chart showing synthetic postal-code regions on the y-axis and mean PLRS for each region on the x-axis. Bars are coloured along a pastel teal-to-red gradient according to average PLRS, with darker outlines and increased border thickness marking regions above an interactive percentile threshold of the mean PLRS distribution. A search box permits filtering to regions whose postal codes match a user-entered substring. Clicking on a bar reveals textual details about the region, including its mean PLRS. This figure mimics a choropleth-style view of polygenic lipid burden across geographical units.

Figure 3. Triage scatter plot linking polygenic lipid risk to Achilles tendon phenotype and family history.
Scatter plot of individual participants with PLRS on the x-axis and Achilles tendon thickness (mm) on the y-axis. Each point’s fill colour encodes PLRS percentile band (pastel teal for low, transitioning through pastel yellow/green and orange to pastel red for the highest bands), while point radius and border width are increased for individuals with a positive family history of premature cardiovascular disease. Interactive controls allow users to filter the display by minimum PLRS percentile and to restrict the view to family-history–positive individuals. Clicking on a point displays its participant ID, PLRS value, percentile, Achilles tendon thickness, family history status, and postal region. This figure illustrates how a dashboard might visually highlight candidates for further monogenic dyslipidemia evaluation or specialist referral by combining polygenic burden, tendon phenotype, and family history in a single triage visualization.