Polygenic risk
A weighted sum of thousands of small-effect variants, compared against a reference distribution. It indicates a relative position within a population, not an individual probability. The conditions are grouped by body system.

All 14 scores, by what they mean for you
Metabolic
2 conditions · all in the average range2 conditions · all in the average rangeDiabetes, obesity, and lipids. This is the group where the modifiable outweighs the genetic — and by a wide margin.
Metabolic
2 conditions · all in the average range2 conditions · all in the average rangeType 2 diabetes
BClear individual signalAbove the average for people of similar genetic ancestry. The weight of the modifiable factors (body weight, physical activity, diet) is still greater than that of the score.
Body mass index
BClear individual signalA somewhat above-average predisposition to gain weight easily. It is the clearest example in the whole report that genetics is not in charge: diet and physical activity explain much more of the real variation in weight than this score does.
Cardiovascular
2 conditions · all in the average range2 conditions · all in the average rangeHeart and blood vessels. The scores contribute most in young people who do not yet have the classic risk factors.
Cardiovascular
2 conditions · all in the average range2 conditions · all in the average rangeCoronary artery disease
BClear individual signalWithin the average. It does not change clinical management: the traditional factors (blood pressure, lipids, smoking) prevail.
Atrial fibrillation
BClear individual signalSlightly above the average, with no clinical significance on its own.
Oncological
2 conditions · all in the average range2 conditions · all in the average rangeNo polygenic score evaluates high-penetrance genes (BRCA, Lynch). If there is a family history, a separate hereditary cancer test is warranted.
Oncological
2 conditions · all in the average range2 conditions · all in the average rangeProstate cancer
BClear individual signalWithin the average. This score does NOT evaluate BRCA2 or other high-penetrance genes: if there is a family history, a separate hereditary cancer test is warranted.
Colorectal cancer
BClear individual signalBelow the average. It does not replace or postpone screening by colonoscopy or occult blood test according to age.
Autoimmune
5 conditions · 1 below percentile 105 conditions · 1 below percentile 10The group with the greatest real genetic weight, mainly because of the HLA region. Here the scores perform better than in the rest.
Autoimmune
5 conditions · 1 below percentile 105 conditions · 1 below percentile 10Celiac disease
BClear individual signalFar below the average, consistent with the negative HLA-DQ2/DQ8 result from the nutrigenetics section. When two independent methods agree, the conclusion is firmer: celiac disease is very unlikely.
Type 1 diabetes
BClear individual signalWell below the average. It is one of the best-performing scores in all of common-disease genomics, because a good part of the risk is concentrated in the HLA region, with large, well-characterized effects. It is used clinically to distinguish type 1 diabetes from type 2 and from the monogenic forms (MODY) in doubtful cases.
Rheumatoid arthritis
BClear individual signalWithin the average. Autoimmunity scores perform better than the rest because a good part of the weight is in the HLA region, which has large, well-characterized effects.
Autoimmune hypothyroidism
BClear individual signalSlightly above the average. It is a common condition, simple to diagnose with a blood test and with well-established treatment: if symptoms appear, it is investigated and resolved.
Inflammatory bowel disease
BClear individual signalBelow the average for Crohn's and ulcerative colitis.
Neurological
2 conditions · all in the average range2 conditions · all in the average rangeIt includes conditions with no prevention available today: knowing the number does not enable any management that improves the prognosis, and for some people knowing it weighs on them. They are reported anyway, with that caveat.
Neurological
2 conditions · all in the average range2 conditions · all in the average rangeMigraine
BClear individual signalWithin the average.
Alzheimer's disease
BClear individual signalRestricted-access result: there is no proven intervention today that modifies this risk, and knowing it can cause distress without offering an action in return. It is reported only if you explicitly requested it, and reviewing it in consultation rather than reading it alone is recommended.
Bone and muscle
1 condition · all in the average range1 condition · all in the average rangeBone density and fracture risk, where preventive management is in fact well established.
Bone and muscle
1 condition · all in the average range1 condition · all in the average rangeOsteoporosis
BClear individual signalSlightly above the average in risk of low bone density. Here prevention is in fact well established and inexpensive: calcium, vitamin D, weight-bearing activity, and bone densitometry according to age.
Not calculated from the panel
The panel carries 16 conditions and 14 were calculated for this sample. The missing ones are listed with the reason: a score that was not calculated is not a low score.
- Breast cancer — only reported in females
- Ovarian cancer — only reported in females
These scores are NOT calculated directly on what the chip measures. Measured against the actual GSA v3 manifest over 60 PGS Catalog scores: the array covers a median of 18% of the variants each score needs —half of the scores fall between 15% and 23%— and none of the 60 reaches the 90% that would be needed to calculate it outright. That is why the imputation step is mandatory, not optional: from the inherited blocks, the missing positions are inferred against a panel of tens of thousands of genomes. A score calculated without imputing would not give an attenuated result, it would give a wrong one — it would end up off-center with respect to the reference distribution it is compared against.
Percentiles will be normalized by continuous genetic ancestry (projection onto principal components), instead of assigning a fixed population label. That calculation depends on the imputation step, which is not yet resolved: without imputing, none of the scores measured reaches the necessary coverage. Even once resolved: most published scores were derived from cohorts of European ancestry, and their performance in admixed Latin American populations is less validated. To be interpreted as one more layer of information, never as a diagnosis nor as an exact individual probability.