Your genetics can tell you something about inherited susceptibility. Your lab results can tell you something different about what is measurable now. Your medical records can show what has actually happened over time.
Your medications, symptoms, diagnoses, imaging, family history, and treatment history add still more context.
Artificial intelligence becomes most useful when these layers are brought together carefully.
A useful health-AI workflow is not:
Upload your DNA and ask AI what is wrong with you.
It is:
Structure the genetic evidence
↓
compare it with current laboratory data
↓
compare it with longitudinal medical history
↓
consider medications, symptoms, and alternative explanations
↓
strengthen, weaken, or refine the genetic hypothesis
That is the direction Mutant Genomics is designed around.
Mutant first converts compatible DNA data into:
Relevant genetic markers
↓
Genetic module patterns
↓
Converging patterns
↓
Ranked health hypotheses
↓
Supporting evidence + DNA coverage
Those findings can then become structured Health Context for use with laboratory results, medical records, medications, symptoms, diagnoses, imaging, family history, supplement history, and other relevant health information.
The useful division of work is:
Mutant interprets the genetics.
AI helps connect it to the rest of your health story.
Your top 3 ranked health hypotheses are included in full with Mutant Free. No credit card required.
Your complete raw DNA stays in your browser. Mutant retains only the genetic markers needed for your analysis.
To combine DNA, labs, and medical records with AI effectively:
The goal is not:
Make everything fit the genetics.
The goal is:
Determine whether the genetics actually help explain the health question.
DNA is unusually powerful because it remains relatively stable throughout life.
But that is also its limitation.
Your DNA may suggest greater susceptibility to a biological pattern without telling you whether that pattern is currently affecting your health.
cannot determine current:
cannot determine current:
cannot determine current:
cannot determine current:
cannot determine current:
cannot determine current brain:
DNA can generate a useful question.
It usually cannot answer the whole question.
Laboratory tests measure something much closer to the current biological state.
That can include:
But a lab result does not always tell you why the result looks that way.
For example, an elevated biomarker could potentially reflect:
A laboratory value is a measurement. It still needs interpretation.
Medical records provide something neither DNA nor a single laboratory result can provide: time.
Records may show:
A health hypothesis that looks plausible from DNA may become much less compelling after reviewing several years of medical history.
The reverse can also happen.
A useful AI-health analysis can think in five major layers.
Genetic data may provide inherited context involving:
This layer asks: what does the inherited evidence make more or less plausible? It does not establish what is happening now.
Laboratory data may show:
This layer asks: is there current biochemical evidence compatible with the genetic hypothesis?
Records may reveal:
This layer asks: does the longitudinal history fit the hypothesis?
Relevant exposures may include:
This layer asks: could something outside the genetics explain the same findings?
Symptoms add the lived biological pattern.
Useful information includes:
This layer asks: does the real-world phenotype resemble the proposed biological mechanism?
Traditional medical information is often fragmented.
You may have:
AI can be useful for:
But the quality of the reasoning depends heavily on how the genetic layer is presented.
A 23andMe or AncestryDNA file may contain hundreds of thousands of genotype calls. Whole-genome sequencing can contain vastly more information.
If you simply provide raw DNA and ask, “What explains my health problems?”, AI first has to determine:
That is a large genomics problem before the medical reasoning even begins.
Mutant reduces the genetic search space first.
The process is:
Raw DNA
↓
Relevant genetic markers
↓
Genetic module patterns
↓
Converging patterns
↓
Ranked health hypotheses
↓
Supporting evidence + DNA coverage
Now AI does not have to begin with “Here are hundreds of thousands of SNPs.”
It can begin with something closer to:
“Here is the biological hypothesis, why the genetics support it, how complete the genetic evidence is, and what real-world evidence should strengthen or weaken it.”
That is a much better AI problem.
Mutant uses Health Context to describe the non-genetic information that can help evaluate a genetic hypothesis.
Conceptually:
hypothesis
├── interpretation
├── evidence
├── dna_coverage
├── correlation_context
│ ├── relevant_labs
│ ├── relevant_symptoms
│ ├── relevant_history
│ ├── medications_exposures
│ ├── strengthens_hypothesis
│ ├── weakens_hypothesis
│ ├── alternative_explanations
│ └── questions_to_explore
└── sources
The key idea is that Health Context should not merely tell AI how to confirm the finding. It should also tell AI how the finding could be wrong, irrelevant, incomplete, or better explained another way.
This is one of the most important principles in AI-assisted health interpretation.
Suppose genetics suggest lower folate-dependent methionine-recycling reserve.
A weak AI workflow searches only for evidence that confirms it. A stronger workflow asks:
Perhaps:
Perhaps:
The correct conclusion may become:
The genetic susceptibility exists, but it may not be particularly relevant to the current health question.
That is a useful result.
Whenever you use AI with genetic Health Context, useful prompts include:
What evidence contradicts this hypothesis?
What alternative explanations fit the same labs and symptoms?
Which findings in my medical records make this hypothesis less likely?
Does medication explain the pattern better?
Which hypothesis has the strongest direct biochemical support?
This helps reduce confirmation bias.
Suppose Mutant identifies lower cellular T3-activation reserve. This might be influenced by genetic patterns involving DIO1, DIO2, and supporting thyroid-related biology.
The genetics do not establish that you currently have poor T4-to-T3 conversion.
Health Context may include:
AI can then ask:
Do the laboratory results support a thyroid-centered explanation? Could illness or calorie restriction explain the T3 pattern? Are the thyroid results repeatedly reassuring? Does another hypothesis fit better?
Suppose Mutant identifies HFE-related iron-overload susceptibility. That finding should not be converted directly into “you have iron overload.”
Health Context can include:
AI can compare genetic susceptibility against actual biochemical evidence. If relevant iron markers remain normal over time, the hypothesis may have less current importance. If compatible abnormalities are present, it may deserve more attention and appropriate clinical confirmation.
Lipoprotein(a) is strongly influenced by genetics. A Mutant hypothesis may suggest greater inherited Lp(a)-related susceptibility.
But the real-world question still requires:
AI can then ask: does the measured Lp(a) support the inherited pattern? This is a good example of DNA generating the question and the laboratory result measuring the current phenotype.
Suppose Mutant identifies a histamine-related hypothesis involving AOC1/DAO-related intestinal clearance, HNMT-related intracellular metabolism, gut motility, and immune reactivity.
The genetics cannot diagnose histamine intolerance.
Useful Health Context may include:
AI can then compare competing explanations: is this primarily food-derived histamine exposure, is gut motility lowering tolerance, does an allergy explanation fit better, or are symptoms unrelated to food?
Suppose Mutant identifies lower methionine-recycling reserve. The genetic evidence may involve MTHFR, MTR, MTRR, BHMT, and supporting B12 and folate biology.
Useful Health Context may include:
AI can help determine whether the genetics are supported by current biochemistry, whether a B12 issue is more important than folate, whether strong compensation exists, and whether the pathway appears adequately supported despite the genotype.
A potentially important inherited finding such as G6PD deficiency requires a different standard from a common pathway SNP.
If relevant genetic evidence appears, Health Context may include:
AI can help organize the evidence. It should not turn an unconfirmed consumer genotype into a definitive diagnosis.
Suppose potentially relevant F5 evidence appears.
Useful Health Context may include:
The AI question becomes: how important is this inherited finding in the person's actual clinical context? Potentially significant findings may require clinical confirmation.
Suppose Mutant identifies enteric motility & clearance susceptibility.
Health Context may include:
AI may discover that genetics + symptoms support a motility hypothesis, or instead that documented pelvic-floor dysfunction explains the constipation better. The genetic hypothesis should be allowed to lose.
Suppose Mutant identifies patterns involving Circadian & Sleep, Catecholamines & Arousal, and Stress & Autonomic Regulation.
Useful Health Context may include:
AI can ask: is the problem more consistent with circadian timing, stimulant exposure, ADHD-related executive arousal, anxiety, thyroid dysfunction, sleep apnea, or another cause?
Suppose someone experiences strong reactions to methylfolate, B12, magnesium, and vitamin D.
A simple genetic interpretation might assign a different pathway problem to each one.
Health Context can instead examine:
AI may conclude that the genetics do not explain the reactions as well as the formulation, dose, or medication context does.
One isolated value can be misleading. AI becomes particularly useful when you provide dates, reference ranges, and repeated measurements.
| Date | TSH | Free T4 | Notes |
|---|---|---|---|
| Jan | 2.1 | 1.2 | No thyroid medication |
| Apr | 2.3 | 1.1 | Symptoms unchanged |
| Aug | 2.0 | 1.2 | Similar pattern |
A longitudinal pattern often tells you more than one isolated result.
The same applies to ferritin, transferrin saturation, LDL, Lp(a), B12, folate, homocysteine, glucose, and blood counts.
Laboratory ranges can differ by laboratory, assay, units, population, and testing method.
Instead of giving AI only “Ferritin = 40”, provide “Ferritin = 40 ng/mL, reference 30–400” when available.
That reduces ambiguity.
A laboratory value without units can be misleading.
For example, “Vitamin B12 = 300” is incomplete. AI should ideally receive value, unit, reference range, and date.
The same is true for thyroid hormones, lipids, iron, glucose, vitamin D, and hormones.
A genetic result from today can be compared against labs from last week or labs from 10 years ago—but those values describe different biological moments.
Medical reasoning improves when AI knows:
Timeline matters.
Instead of giving AI hundreds of pages with no structure, consider organizing major events such as:
2022
- Fatigue begins
- TSH 2.3, Free T4 1.1
2023
- Constipation worsens
- Started medication X
2024
- Ferritin 18
- Began iron treatment
2025
- Ferritin 55
- Fatigue persists
2026
- Mutant identifies thyroid and sleep-related hypotheses
AI can still use the source records. But the timeline creates a much clearer reasoning framework.
Suppose symptoms began immediately after starting a stimulant, increasing thyroid medication, beginning an antidepressant, starting iron, taking an antihistamine, or starting a GLP-1 medication.
That timing may be more informative than a genetic susceptibility identified years later.
AI should ask: did the phenotype begin before or after the exposure?
A plausible biological pathway is not enough. Chronology matters.
People often collect only abnormal test results. That can bias AI toward finding disease.
Normal or reassuring results can be extremely valuable because they may weaken a hypothesis.
For example, MTHFR susceptibility plus normal folate + normal B12 + normal homocysteine is different from MTHFR susceptibility + low folate + elevated homocysteine.
Likewise, thyroid susceptibility + repeated reassuring thyroid tests may make the genetic hypothesis much less relevant to a current symptom.
Do not omit normal data simply because it looks uninteresting.
Medical records may show that a suspected condition has already been evaluated.
Examples:
These findings may weaken one hypothesis, redirect attention toward another, or prevent AI from repeatedly suggesting questions that have already been investigated.
One of the strongest uses of AI is finding things humans overlook when information is scattered.
For example:
A diagnosis in one note says iron deficiency, but repeated iron studies may not support it.
Several notes attribute symptoms to thyroid disease, but the symptom onset predates the thyroid abnormality.
A genetic hypothesis suggests histamine-related food reactivity, but the record shows episodes occurring while fasting.
These contradictions help improve reasoning.
Imaging can add another Health Context layer.
Examples may include:
AI can help summarize whether an imaging finding supports the hypothesis, provides a competing explanation, or is unrelated.
Genetics generally cannot substitute for direct structural information.
Wearable data can sometimes add useful context involving heart rate, resting heart rate, exercise patterns, sleep timing, activity, and recovery trends.
But wearable estimates are not always equivalent to clinical measurements. AI should treat them as an additional context layer rather than a diagnostic replacement.
Family history can substantially change how inherited genetic findings are interpreted.
Examples include:
Useful details include relationship, condition, approximate age at diagnosis, and age at major event when known.
Family history can make a genetic finding more important. It can also identify questions that common consumer DNA does not fully answer.
Symptoms should be described precisely.
Instead of “food intolerance”, describe:
Instead of “anxiety”, describe:
Precise phenotype gives AI better information.
Diagnostic labels may be uncertain. A symptom pattern is often safer and more informative.
For example:
Heart pounding + sweating + tremor during social stress, lasting 90 minutes after the event
contains more information than “I have dysautonomia.”
Similarly:
Bloating increases through the day and worsens after several days without a complete bowel movement
contains more useful motility context than “I have food intolerance.”
When combining health data, it helps to separate:
from:
Keeping those layers separate reduces accidental certainty.
Instead of asking, “Do I have a methylation problem?”, ask:
What specific evidence supports the methionine-recycling hypothesis, what evidence argues against it, and which facts are still missing?
Instead of asking, “Is my thyroid genetic problem causing constipation?”, ask:
Compare the thyroid and gut-motility hypotheses using my labs, bowel history, medication history, and genetic evidence.
This makes the reasoning easier to inspect.
Not all evidence deserves equal weight.
A useful hierarchy may distinguish:
AI should not automatically weigh them equally.
Consider someone with clear inherited HFE susceptibility but normal transferrin saturation and ferritin.
The genetic finding may still matter for future awareness. But it may not explain current fatigue.
That distinction prevents inherited risk from swallowing the entire health narrative.
The reverse is also true.
Suppose someone has low vitamin B12 but no obvious genetic susceptibility in Mutant's modeled B12 pathways.
The low B12 still matters. Possible causes could include:
Genetics is not required for a real biological problem to exist.
If medical records show documented disease, clear laboratory abnormality, a structural problem, or a medication effect, that evidence may matter more than a modest common genetic susceptibility.
Mutant's purpose is not to replace those findings. It is to add a structured inherited layer.
When using structured Mutant findings with labs and medical records, a strong prompt can ask AI to:
Evaluate this genetic health hypothesis against the supplied Health Context. Distinguish genetic susceptibility from current disease. Identify evidence that strengthens the hypothesis, evidence that weakens it, competing explanations, important missing information, and any finding that may warrant clinical confirmation. Do not assume the hypothesis is correct simply because the genetics support it.
This creates a much better reasoning task than “Tell me what I have.”
Be cautious with prompts such as:
The available evidence usually cannot support that level of precision.
Some genetic platforms calculate polygenic risk scores for selected diseases. Mutant uses a different framework for many of its findings.
Mutant evaluates:
This is designed to help answer which biological mechanisms deserve attention rather than assigning a disease probability to every symptom.
Even when genetic evidence, symptoms, labs, and medical history all appear compatible, the correct conclusion may still be “this hypothesis has stronger support”.
That is not always the same as “this condition is diagnosed.”
Appropriate diagnosis may require:
Mutant evaluates selected findings within Clinical & Inherited Health, including Cardiometabolic & Vascular Health and Blood, Iron & Laboratory Patterns.
Potentially important inherited findings may involve areas such as:
These deserve different treatment from common susceptibility SNPs. A potentially significant finding may require clinical confirmation.
Health Context can also help evaluate findings across:
Each hypothesis can require different Health Context. There is no universal lab panel that validates every genetic finding.
Consider fatigue.
Possible Mutant hypotheses could involve:
AI can then compare those hypotheses against CBC, ferritin, iron studies, B12, folate, thyroid testing, sleep history, medication, and medical records.
The useful question is not “Which fatigue gene do I have?”. It is “Which hypothesis has the strongest combined evidence?”
Systems do not operate independently.
For example, reduced thyroid function may contribute to slower gut motility, which may contribute to more fermentation and food sensitivity, which may increase interest in histamine-related symptoms. Meanwhile, poor sleep may increase stress and autonomic reactivity.
A static SNP report may list these as unrelated findings. Health Context lets AI examine whether a cross-system relationship actually fits the person's history.
Two people can carry similar genetic patterns.
Person A may have relevant symptoms, supporting labs, and compatible medical history. Person B may have normal labs, no relevant symptoms, and strong evidence for another explanation.
The genetics may be similar. Their Health Context is different. Therefore, the practical interpretation should also be different.
Unlike DNA, Health Context is dynamic.
It can change with age, medication, illness, diet, pregnancy, menopause, weight change, surgery, infection, sleep, exercise, stress, and treatment.
A genetic hypothesis that was irrelevant five years ago may become more interesting later. Or the opposite.
Suppose Mutant identifies a thyroid-related hypothesis today, and current testing is reassuring.
The conclusion may be: genetic susceptibility exists, but current evidence does not strongly support thyroid dysfunction.
Years later, new symptoms, labs, and medications may change the Health Context. The genetics did not change. Their relevance did.
Combining genetics, medical records, labs, medication, and family history creates a particularly sensitive dataset.
Before sharing information with any AI platform, understand:
Do not assume every AI platform handles health information the same way.
Mutant does not require your complete raw DNA file to be uploaded and stored.
When you add compatible DNA data:
Your complete raw DNA is not uploaded or stored by Mutant.
Once the genetic interpretation has been structured, AI can work with:
That allows the AI conversation to focus on the health question rather than the entire genome.
Medical records may contain name, address, date of birth, phone number, email, medical-record number, insurance information, employer, and other identifying details.
If you are manually providing records to an AI tool, consider whether those fields are necessary. For many health questions, they are not.
You may be able to remove identifying information while retaining diagnoses, labs, imaging, medication, and clinical notes. Use the privacy controls and policies appropriate to the AI service you are using.
Genetic and medical records may also contain health information about parents, children, siblings, and other relatives.
That information may help with inherited-risk interpretation. It also deserves privacy consideration.
If you are evaluating an HFE/iron hypothesis, the relevant information may be the HFE finding, ferritin, transferrin saturation, CBC, liver context, and family history.
You may not need years of unrelated dermatology notes, every unrelated genotype, or an entire genome file.
Focused context can improve both privacy and reasoning quality.
A data dump gives AI everything. Structured Health Context gives AI what is relevant, why it is relevant, and how it should be challenged.
That distinction matters.
A useful export may include:
That gives another AI tool enough context to reason without requiring it to recreate Mutant's entire genetic-analysis pipeline.
Structured formats can make it easier for AI to distinguish hypothesis, evidence, limitations, labs, symptoms, and alternative explanations from one another.
For example:
hypothesis:
name: Lower cellular T3 activation reserve
interpretation:
summary: >
Genetic evidence suggests lower inherited reserve in pathways
involved in thyroid hormone activation.
health_context:
relevant_labs:
- TSH
- Free T4
- T3 when clinically relevant
strengthens_hypothesis:
- compatible thyroid laboratory pattern
weakens_hypothesis:
- repeated reassuring thyroid testing
alternative_explanations:
- illness
- calorie restriction
- medication effects
This is much more useful than an unstructured list of SNPs.
A good AI response should use language such as supports, weakens, is compatible with, raises the possibility, and does not establish—rather than proves, confirms, or your genes caused when the evidence does not support that level of certainty.
Suppose Mutant wants ten markers for a module but only six are present in the DNA source.
That should remain part of the interpretation. AI should understand: there is evidence, but coverage is incomplete. It should not silently assume the missing markers are normal.
Mutant supports compatible data from 23andMe, AncestryDNA, and selected whole-genome sequencing (WGS) formats.
The same underlying analysis framework applies. What changes is how much relevant genetic evidence is available.
23andMe may provide useful common genetic markers across many Mutant models. Coverage varies by testing version. You can use compatible existing data without purchasing a new DNA test.
Compatible AncestryDNA files can also provide useful genetic coverage. Some markers available in 23andMe may be missing in AncestryDNA and vice versa. Missing remains missing.
WGS generally provides broader coverage. That may mean fewer missing modeled variants, more complete modules, better evidence coverage, and less uncertainty for some hypotheses.
But Mutant does not interpret every variant in an entire genome. It evaluates the subset incorporated into its current models.
If broader DNA becomes available, an existing hypothesis may strengthen, weaken, reveal compensation, stay largely unchanged, or lose rank to another hypothesis.
The purpose of more complete DNA is better evidence—not more things wrong.
More medical data should not automatically make a genetic hypothesis stronger.
Additional records may reveal reassuring testing, a competing diagnosis, medication effects, or a different timeline.
More context should increase accuracy—not confirmation.
Mutant's Health Context direction is not:
The role is to help AI compare evidence, organize information, surface contradictions, evaluate alternative explanations, identify missing context, explain genetic findings, and distinguish inherited susceptibility from current state.
Mutant does not diagnose hypothyroidism, histamine intolerance, MCAS, IBS, SIBO, ADHD, anxiety disorders, depression, nutrient deficiencies, neurotransmitter abnormalities, or most complex chronic conditions.
Mutant provides educational genetic context and ranked health hypotheses.
Selected Mutant findings may point toward inherited patterns with greater potential clinical relevance.
A consumer-DNA finding may still require clinical laboratory confirmation, appropriate biomarker testing, family-history review, specialist evaluation, and genetic counseling.
AI can help explain why confirmation matters. It should not skip that step.
Genetic counselors and clinical genetics specialists can provide testing strategy, family-history assessment, variant interpretation, inheritance counseling, reproductive counseling, and clinical implications.
AI can help users understand information. It does not replace these professional functions.
Likewise, combining DNA, labs, and medical records with AI does not replace physical examination, diagnostic testing, clinical judgment, emergency evaluation, or treatment planning.
The value is in better organization and hypothesis testing.
If you want to combine your own DNA and health information with AI, a useful sequence is:
Use compatible 23andMe, AncestryDNA, or selected WGS data. Your complete raw DNA is processed locally in your browser.
Do not begin with every variant. Begin with: which biological hypotheses actually have the strongest genetic support?
For each important hypothesis, review contributing patterns, supporting variants, evidence, missing DNA, and limitations.
Ask which labs matter, which symptoms matter, which diagnoses matter, which medications matter, and which history would weaken it.
Prioritize dates, values, units, reference ranges, and trends—rather than simply collecting every laboratory result ever performed.
Focus on diagnoses, important negative workups, imaging, procedures, family history, specialist conclusions, and major changes over time.
Include name, dose, timing, start date, stop date, and major dose changes when known.
Include symptom onset, triggers, timing, progression, what improves it, and what worsens it.
Remove identifying information that is not needed for the analysis if you are manually sharing records with an AI service.
Do not ask only “Does this prove the hypothesis?” Ask what supports it, what weakens it, what alternative explanation fits better, and what information is still missing.
Lower cellular T3-activation reserve.
Moderate. Several thyroid-activation-related module patterns contribute.
Partial. Some desired markers were not available from the DNA source.
The genetics support a thyroid-activation susceptibility, but the available clinical evidence does not establish that this is the primary cause of the current symptoms.
That is much more useful than “You have a DIO2 mutation.”
This may be one of the highest-value outcomes.
For example:
The genetic histamine hypothesis is supported, but your records show that symptoms began immediately after a medication change and do not correlate with food exposure. The medication timeline may be more relevant to the current problem.
Your MTHFR-related susceptibility is real, but folate, B12, and homocysteine are reassuring. It may not explain your current fatigue.
The thyroid hypothesis is genetically plausible, but a documented sleep disorder provides stronger evidence for daytime fatigue.
That is what evidence-based personalization should look like.
Mutant's job is to answer: what does the genetic evidence support?
AI Health Context helps answer: does the rest of the person's health information make that hypothesis more or less relevant?
Those are separate tasks. Keeping them separate helps avoid SNP overinterpretation, confirmation bias, genetic determinism, unnecessary supplement protocols, and diagnosing from common variants.
You do not need to complete a Mutant questionnaire to generate your genetic analysis. Health Context is optional.
It can come from Mutant's optional questionnaire, laboratory results, medical records, medication history, symptoms, diagnoses, family history, and other health information.
This is particularly useful for people who plan to explore their Mutant findings with AI.
You do not need to purchase a new genetic test to begin.
If you already have compatible 23andMe or AncestryDNA data, Mutant can analyze the relevant markers available from that source. You can later decide whether broader DNA coverage would actually add value.
Mutant Free includes:
Your top 3 hypotheses can come from any biological system or hub.
No credit card required.
Mutant Full unlocks all remaining findings available from your analysis.
Included:
You can start free and upgrade later.
Mutant Free does not run a smaller genetic analysis. The same underlying framework generates the findings.
Your top 3 ranked health hypotheses are fully accessible.
All remaining ranked hypotheses are also unlocked.
A free hypothesis is not a teaser. You can review explanation, score, contributing patterns, supporting evidence, DNA coverage, Health Context, and limitations in full.
Mutant is designed so your structured genetic findings can be used beyond the Mutant portal.
Instead of exporting a giant list of raw SNPs, an AI-ready export can preserve hypotheses, genetic evidence, contributing patterns, coverage, limitations, relevant Health Context, evidence that strengthens or weakens the hypothesis, and alternative explanations.
That gives another AI tool a much more useful genetic layer to work from.
Mutant's structured Health Context is designed for AI-assisted reasoning across genetics, labs, medical records, medications, symptoms, and history.
Availability of direct integrations may depend on the specific supported AI experience. You can also use Mutant's AI-ready export in compatible AI workflows.
Mutant does not upload or store your complete raw DNA file.
When you add compatible DNA data:
Yes. This can be particularly useful when genetics suggest susceptibility and laboratory testing measures the current biological state. The lab result may strengthen or weaken the genetic hypothesis.
Yes. Medical records add diagnoses, lab trends, imaging, medication history, specialist findings, and prior testing that DNA cannot provide.
AI can reason across genetic and laboratory information. The genetic interpretation is more reliable when the raw DNA has first been structured into meaningful evidence rather than asking AI to infer everything directly from a giant raw genotype file.
Yes. Longitudinal trends may be more informative than a single result. Include dates, units, and reference ranges when possible.
Many modern AI systems can work with uploaded documents and extract information from them. Accuracy should still be checked, especially for units, reference ranges, dates, and unusual formatting.
Yes. Normal findings may weaken a hypothesis and are therefore important evidence. Providing only abnormal results can create confirmation bias.
Yes. AI can be especially useful for identifying when symptoms began, when medication changed, how labs evolved, and whether an abnormality was temporary or persistent.
Yes. A structured approach can ask which hypothesis has stronger genetic evidence, stronger laboratory support, better phenotype fit, or more contradictory evidence.
AI can compare evidence. It cannot always establish that one hypothesis is definitively correct. Some questions require clinical testing.
Combining multiple data sources can make a hypothesis more or less plausible. That still does not automatically establish a medical diagnosis. Diagnosis may require specific criteria, testing, examination, or clinical confirmation.
It can help compare the genetic finding against folate, B12, homocysteine, CBC, diet, and other genetic compensation. The conclusion may be that the variant has limited current relevance.
It can compare a thyroid-related genetic hypothesis against TSH, Free T4, T3 when relevant, thyroid antibodies, medication, and medical history. Genetics alone cannot determine current thyroid function.
Not from genetics alone. Useful context may include food freshness, meal timing, constipation, DAO response, allergy history, hives, and symptoms occurring without food.
No single genetic result or AI analysis definitively establishes histamine intolerance. AI can help compare the hypothesis against broader Health Context.
No. MCAS requires appropriate clinical evaluation and cannot be diagnosed from DAO, HNMT, MTHFR, or common immune-related SNPs.
No. HFE genetics indicate inherited susceptibility. Current iron status depends on measures such as transferrin saturation, ferritin, and other clinical context.
Yes. Genetics can provide inherited context while actual Lp(a) and lipid measurements provide current measurable evidence.
AI can help organize the genetic finding, personal clotting history, family history, medication, hormonal exposure, and surgery history. Potentially important inherited findings may still require clinical confirmation.
It may help identify questions worth discussing. But supplement need and dose depend on actual deficiency, health conditions, medication, kidney function, liver function, diet, and safety. Mutant does not create supplement prescriptions from common SNPs.
General pathway genetics should not be used for medication selection. Specialized pharmacogenetic testing is a different use case. Mutant is not currently positioned as a medication-selection platform.
Yes. This can be useful for asking whether medication provides a stronger explanation for symptoms, lab changes, sleep changes, or gastrointestinal changes than the genetic hypothesis does.
Yes. Family history can make certain inherited findings more or less important.
Yes. Imaging can provide structural Health Context that genetics cannot provide. Important findings should still be interpreted clinically.
Potentially. Wearable information can add context involving sleep timing, heart rate, activity, and exercise, but should not automatically be treated as equivalent to clinical testing.
Not necessarily. Focused context may improve privacy, relevance, and reasoning quality. Provide the information relevant to the health question when practical.
If identifying information is not needed for the task, de-identifying manually shared medical records can reduce unnecessary exposure of personal information. Review the privacy policies and settings of the AI service you use.
You can use AI services according to their current file and privacy policies. But you should understand how the platform handles genetic information before uploading a complete genome or raw DNA file. Mutant avoids requiring your complete raw DNA file to be stored on its servers.
No. Your browser reads the file locally. Only genetic markers required for Mutant's current models are sent and retained.
No. Unavailable evidence remains missing.
Yes. Compatible 23andMe data can provide useful coverage across many Mutant models.
Yes. Compatible AncestryDNA data is supported.
Yes, selected compatible WGS formats are supported. WGS typically provides broader marker coverage.
No. You can begin with compatible consumer DNA data you already have.
More complete DNA can reduce uncertainty. But additional evidence may strengthen or weaken a hypothesis.
No. Mutant provides educational genetic analysis and structured Health Context. Its findings are hypotheses rather than diagnoses.
Mutant provides a structured genetic-analysis layer designed to work well with AI. Mutant focuses on genetic markers, modules, converging patterns, ranked hypotheses, evidence, and coverage. AI can then help compare those findings against broader health information.
Yes. Mutant's AI-ready export is designed to make structured genetic Health Context portable to compatible AI workflows.
No. The questionnaire is optional. You can use other Health Context such as labs, medical records, medications, symptoms, diagnoses, and family history.
Mutant Free includes your top 3 ranked health hypotheses across your entire analysis in full. Your top three can come from any biological system or hub.
No. Free controls breadth, not depth. Your included hypotheses are fully accessible.
Mutant Full is $49/year. It unlocks all remaining ranked hypotheses available from your analysis.
The most useful AI-health model does not force every data source to do the same job.
DNA asks: What biological susceptibilities did I inherit?
Labs ask: What can be measured now?
Medical records ask: What has actually happened over time?
Medications ask: What external factors may be changing the biology?
Symptoms ask: What is the person actually experiencing?
AI asks: How well do these layers fit together—and where do they disagree?
Mutant provides the structured genetic layer:
Relevant genetic markers
↓
Genetic module patterns
↓
Converging patterns
↓
Ranked health hypotheses
↓
Supporting evidence + DNA coverage
↓
Health Context
AI can then help determine what supports the hypothesis, what weakens it, what alternative explanation fits better, which information is still missing, and which potentially inherited finding deserves clinical confirmation.
The objective is not to make the genetics explain everything. It is to determine whether the genetics are actually relevant to the current health question.
Mutant interprets the genetics.
AI helps connect it to the rest of your health story.
Your top 3 ranked health hypotheses are included in full with Mutant Free.
No new DNA test required. No credit card required.
Your complete raw DNA stays in your browser.
Mutant provides educational and informational genetic analysis. Mutant findings are health hypotheses, not diagnoses, and are not a substitute for medical evaluation, clinical genetic testing, genetic counseling, medication management, or treatment. Potentially significant inherited findings may require confirmation through an appropriate clinical laboratory.