Methodology
How your Precision Longevity Analysis is built.
Here is exactly how your DNA becomes a few clear, practical changes. The evidence we read against, the steps we follow, and the limits we stay honest about, all in plain language. Your genome is not a place for hidden machinery.
Pipeline
Six steps, from your DNA file to your plan.
No black box. We run every one of these steps, in this order, for every analysis. The step names are technical because the work is; what you read at the end is not.
- 01
Quality control
First, we make sure your DNA file is trustworthy. We check that it read cleanly and consistently across the genome, the kind of routine integrity tests a lab runs before anyone draws a conclusion. A file that fails is flagged and re-requested, never read against shaky data.
- 02
Strand harmonisation
Next, we make sure files from different companies speak the same language. 23andMe, AncestryDNA, and others can record the same spot in your DNA in two different ways. We line them all up to one standard reference so the same variant is always recognized as the same variant.
- 03
Imputation
Then we fill careful gaps. Your DNA file reads only certain spots; sometimes an area you've chosen needs a nearby spot that wasn't directly measured. We can infer it from large population data, but only the high-confidence calls, and anything inferred is clearly labeled as such, never passed off as directly read.
- 04
Interpretation
Now a scientist examines the variants that matter, one by one, against the current research. Where ACMG criteria are published for a variant, we cite them. We do not issue our own ACMG classifications, which is the work of clinical labs. And where a finding depends on your ancestry, we say so: a score built in one population is not quietly applied to another.
- 05
Drafting
Your analysis is then written, in plain prose, by the same scientist who read your DNA. The person who understands your findings is the person who explains them to you.
- 06
Review
Finally, everything is read once more, and anything worth taking to a doctor is checked again against its source before it goes out. Your analysis is signed by the scientist who wrote it, and anything of clinical consequence is flagged for you to bring to your physician.
Evidence
What we read against.
Variant classification standards (consulted, not issued)
ACMG / AMP
The medical-genetics rulebook for how serious a DNA change is. Its 2015 standards define how clinical labs grade variants, Pathogenic, Likely Pathogenic, Uncertain Significance, Likely Benign, Benign. We read against these criteria for our own carefulness, but we do not issue our own ACMG classifications: that is the work of clinical-grade labs.
Variant–condition database
ClinVar
The public record, kept by the US National Institutes of Health, of which DNA changes have been linked to which conditions. One change can carry a dozen submissions from a dozen laboratories, and they disagree with each other. We read all of them, ranked by how carefully each was reviewed, not only the alarming ones. On one recent analysis, 53 changes looked worth writing up when only the alarming submissions were counted; two survived reading the rest.
Pharmacogenomic guidelines
PharmGKB / CPIC
The expert guidance on how your genes can affect the way you respond to medications, used for the pharmacology area. PharmGKB gathers the evidence; CPIC turns it into clear prescribing guidance. We follow only the well-supported recommendations and name the studies they rest on.
Primary evidence
PubMed / peer-reviewed literature
The published research itself, the foundation under everything above. We don't write a single finding without naming the papers it comes from, and every citation is listed at the back of your analysis.
Population reference
1000 Genomes · gnomAD
Large catalogues of how common each DNA variant is across the world's populations. A variant means something different when it is rare than when it is common, and the global average is often the wrong number to read yours against: one change we looked at is present in roughly 1 person in 10,000 worldwide, and 1 in 350 in the population that client came from. We read frequency in your own population, and we say which one we used.
The limits
What this analysis cannot do.
A good analysis is defined as much by what it refuses to claim as by what it tells you. Here are the four lines we won’t cross, and why.
We inform, we don't diagnose
This is an informational and educational analysis, not a clinical or diagnostic test. We read what your DNA suggests and write it in plain language; anything of clinical consequence is flagged for you to take to your physician or a genetic counselor. We do not diagnose, treat, or tell you what will happen.
Polygenic scores are directional
Some traits are shaped by thousands of tiny genetic effects added together. We can read which way yours lean, but not predict your future. So we use these scores as direction and context, never as a personal forecast.
Penetrance is not destiny
Carrying a variant, even a serious one, doesn't mean you'll develop the condition. Your other genes, your environment, and plain chance all weigh in. We write findings with that nuance kept in, not stripped out.
The literature moves
Your analysis reflects what the science says on the day it's written, and the science changes. If a variant we wrote about is later reclassified, you're welcome to write to us and we'll send an updated note.
A depth you can choose
When you want to read the whole thing.
Some of the most important questions, rare variants, larger structural changes in your DNA, the fuller drug-response detail, live outside the slice an array covers. Whole-genome sequencing covers your whole genome about thirty times over, reaching what an array physically cannot see.
One scientist goes through your results and signs the analysis at either depth. The same eight areas are written at both, in the same way. What the whole genome adds is depth underneath them. You choose the depth when you apply.
Under method
A Precision Longevity Analysis you can audit, page by page.
Every line in your analysis comes with the variants behind it, the studies they were drawn from, and the date they were last read. Open it, follow it, question it.