On a dark February morning, I wake grainy with sleep and head to the kitchen. Before making toast or coffee, I unscrew the cap from a tiny test tube and spit into it. Over and over, but it’s surprisingly difficult to fill up a whole vial. It takes 10 minutes before my frothy deposit reaches the marked minimum line.

My housemate sips her coffee. “Are you ill?” she asks.

“No,” I say, as I put the pot into a plastic envelope marked “hazard”. “But maybe I could be more well?”

I am healthy. I’m 30, I go to the gym two or three times a week, I eat plenty of vegetables, and I’m lucky to have never suffered from a serious illness. But I also have this nagging sense that I could be fitter, stronger, more resilient to colds. I could be less tired, less prone to mental fogginess. I could be more optimal.

After breakfast, I take my saliva sample to the postbox at the end of my road and send it to 23andMe for analysis. It’s the first of many bodily fluids I’ll be putting in the post to the genetics-testing company over the next six months. In the past few years, dozens of startups have capitalised on our physical anxieties, offering tests that bypass the NHS. They provide pay-as-you-go medical testing normally reserved for the unwell, selling a promise that information equals control.

Self-testing kits are marketed at me everywhere I go. In a train station I walk past a big pink Medichecks stall (“A better you. It’s in your blood”) offering on-the-spot blood tests for everything from cancer markers to thyroid function; on Instagram, Atlas Biomed (“Two tests to discover yourself”) wants to sell me DNA tests and map my microbiome.

These buzzy new companies can broadly be grouped into three: blood tests to analyse health markers (Thriva, LiveSmart, Forth); DNA tests that assess the “fitness potential” in your genes (FitnessGenes, DNAFit); and microbiome analysis that looks at the individual makeup of your gut (Viome, Carbiotix, Atlas Biomed). Each makes broadly similar claims: in exchange for anywhere between £59 and £300, along with a sample of blood, spit or faeces, they will divine the inner workings of your body and advise you on how to improve them. All appropriate the visual language of social media, using pastel hues or young, vital-looking models, or both. This is as much health-for-the-Instagram-generation as a vegan, sugar-free, avocado-chocolate cake. But should you want a slice?

***

Ten weeks after I send off my vial, my 23andMe results are in. The results, it turns out, are innocuous enough, other than the fact that I’m predisposed to age-related macular degeneration. I have to Google the term; it means I may go blind when I’m older. I call my mother. “Your great-grandmother went blind,” she says, before our conversation moves on. Maybe she’s right to be blase: there’s nothing I can actually do with this information at this point. Besides, in the list of possible genetic conditions (such as Alzheimer’s disease, for which 23andMe tells me I have no elevated risk), I feel as if I’ve got off lightly.

A few weeks later, though, the US Food & Drug Administration authorises 23andMe – an American company – as a provider of direct-to-consumer testing for BRCA1/BRCA2 gene mutations, which have been linked to breast cancer. (Once you’ve sent off an initial saliva sample, and your DNA has been sequenced, 23andMe continue to add results to your online profile.) I receive an email from the company: would I like to know if I have the cancer-causing variants? Rationally, I know that a genetic test isn’t destiny. But I also know that the results could open a Pandora’s box, the consequences of which I’m not sure I’m willing to deal with right now. I log off without clicking through on the link.

Am I ready for all this biodata? Last year a study published in the American Journal of Consulting and Clinical Psychology found that telling someone they’re genetically at risk of depression causes them to think of themselves as more depressed. Simply being told that you’re more (or less) at risk of a given condition can have a powerful distorting effect.

According to Avi Lasarow, the CEO and founder of DNAFit – another of the tests I’m about to try – burying my head in the sand puts me in the minority. “There’s been a huge shift in the world of personalisation and proactive health monitoring,” he tells me. “Fitbit made us all more comfortable with tracking our fitness. It marketed the 10,000 steps so well that, before long, we were all comparing our results. And now many of us are ready for the next phase.”

A week later, I go to my gym and am offered a genetic assessment. It’s an expensive, boutique gym popular with young professionals. Co-founder Dave Thomas tells me about his results, one morning after class: “According to my gene types, I’ll respond better if I work out in the afternoons or evenings, and if I eat more carbs. I gave it three months, put a few small tweaks in place, and it has made a noticeable difference.”

I speak to Steven Diffey, a mobile app project manager, who took a 23andMe test two years ago to find out his genetic fitness profile. “It was just curiosity, really,” the 35-year-old tells me. He’d taken up running at 27 and realised he had a natural propensity for it. “I thought I had a bit of potential, so I got a coach and committed to trying to get to the world triathlon championships.” He started a robust regime that included two workouts each day (swimming, cycling, weights, running) and kept a log of everything from his heart rate to how many watts of power he could push with a single downward stroke of the bike pedal. He saw the DNA test as “just another data point that could help with training”.

Diffey’s results changed his race-day mindset. Most world-class sprinters possess the variant of the ACTN3 genotype, which boosts the number of fast-twitch muscle fibres they have (used in powerful bursts of movement). He discovered that he has this variant – something he’d had a hunch about. “Quite often when I race, I still have the ability to sprint at the end. It gives me a little mental boost, though, knowing that my body is perfectly capable of a final burst of speed. I just need to push.”

Talking to Diffey, I get excited about my own DNAFit results, which arrive by email. I’d sent off a cheek swab about 10 days previously – certainly the easiest of the samples I take. I log into my account with a now-familiar twinge of excitement. “Welcome to your genetic journey,” it says. There is a dizzying number of results under the headings of fitness, nutrition and infographic summary. Within fitness there are six tabs, which cover factors such as “training intensity response” and “injury predisposition”. Each gene has a little paragraph of information. I read and reread, but eventually get too frustrated and give up. I don’t understand what “training intensity response” or “reduced toxification ability” means.

It takes an hour-long consultation with DNAFit’s head of product, Andrew Steele, to properly interpret my results. It turns out I have very little sprinting ability, but that I’m “quite rare” in another way. In fact, less than 1% of the DNAFit database – which is tens of thousands – has an “endurance response” as high as I do. I’m practically an Ethiopian distance runner. “It just means that when it comes to reaching your training goal – whether that’s to become faster or strong, to run further, or just to lose weight – you’ll reach that goal quicker if you tailor your training towards endurance.” Instead of 30-second rounds of high-intensity interval training, for instance, he prescribes longer, two- to four-minute rounds, where the pace and intensity is slightly more moderate, but sustained over a longer period.

Apparently I’m also a “low responder” to carbohydrates and saturated fat. In terms of carbs, this means that they don’t cause my blood sugar to spike as significantly as someone who is a “high responder”. That night I load up on sweet potato fries and wholegrain bread. “I’m a low responder to carbs,” I tell my housemate. “So this is fine.”

It takes three weeks (and gaining 5lb) to realise that it’s not. By this point, I have also spoken to two independent specialists in human performance and genetics, both adamant that the science isn’t currently sophisticated enough for us to base either training or nutrition on genetics. “[Genetic testing] companies are snake-oil merchants,” says Hugh Montgomery, a professor of intensive care medicine at University College London, who also directs the Centre for Human Health and Performance (he discovered the first gene to be associated with human performance). “Your genetic code is 6bn letters and it’s carved up into roughly 20,000 genes, which are all interacting and producing proteins. It’s impossible to say exactly how many of them actually influence your training response, because at the moment we don’t know.”

Dr Alun G Williams, director of the sports genomics laboratory at Manchester Metropolitan University, likens it to trying to read a book when you can see only four words on every page. “Most traits to do with fitness are 50% genetic and 50% environment. But we simply can’t test the hundreds of genes that add up to that 50% genetic component, because we only know a small handful of them.”

Even though the ACTN3 genotype – the one that Diffey has and I don’t – has been consistently shown to be related to a person’s sprinting ability, it probably accounts for only a 1-2% difference in how good a sprinter someone is. “Perhaps at an elite level, where an athlete is already running 100m in under 11 seconds, taking that 2% into consideration might be useful,” Williams says. “But for the average consumer who just wants to get fit, the advice isn’t worth the paper it’s printed on.”

A study published in March in the journal Nature found that up to 40% of the analyses of genetic disorders in some at-home testing kits were inaccurate. The study didn’t cover fitness and only looked at 49 samples, so was relatively small; but the advice was that any results had to be backed up by more rigorous testing. It echoes the advice offered by the NHS: an at-home test might be a starting point, but it can’t be used as a diagnostic tool.

***

It takes me three attempts to get enough blood from my fingertips to fill the vial sent to me in the post by Thriva. The company markets itself as a “preventive health service”: it offers regular blood tests to identify deficiencies and signs of possible health problems. I use a small spring-loaded lancet to prick the third finger on my left hand, then dangle it over the sample collection tube, massaging the blood out with my other hand. It forms a few fat, scarlet droplets and then promptly heals. After two more pricks, I have just enough. I’m slightly bruised, but calm; I’m getting used to giving samples now. Like the others, it goes into an airtight envelope before I post it off for analysis.

The three British founders of Thriva are all in their late 20s or early 30s, and come from a tech background. I tell 32-year-old CEO Hamish Grierson I’m not sure how much to extrapolate from my results because I feel, well, fine. “‘Fine’ is a pretty nebulous concept,” he says affably. “We often see people who have been living without feeling as if there’s anything seriously wrong. They tolerate their day-to-day, but they could be feeling considerably better.” Thriva has recently partnered with Vita Mojo, a London restaurant and food delivery service that creates bespoke and personalised meals, based on data from your Thriva results or DNA profile. Low in iron? Have a steak! Need some folate? Here are your leafy greens. Vita Mojo co-founder Nick Popovici likens the service to on-demand TV. “Before Netflix came along, we all watched TV and didn’t realise we wanted something more bespoke. Now, we barely ever sit down at a time that someone else has prescribed. With food it’s the same: why eat the same as everyone else?” He has regular blood tests and takes a host of supplements every day, as well as following a diet tailored to his results.

For Grierson, Thriva’s mission is one of empowerment. He sees a future in which questions such as, “Why am I always tired?” and, “Why can’t we conceive?” can be answered quickly, without necessitating a trip to the doctor. And beyond even that, he envisages a future in which we might never have to ask the question. “We have machine learning models that are able to use our existing data to make a predictive assessment. It’ll then send you a message: ‘Hey, Alex, did you know that, based on our data, people like you are at increased risk of… erm,” he casts around for an example. “Heart attack. Or, you know, diabetes. We’re not there yet – the data set isn’t big enough – but we will be in the next 18 months.

“No one has a big database of bio-grade information on a theoretically ‘well’ population. There is a huge amount of value in it for researchers, who can use it to zero in on the health of various geographies or ethnic groups.” It’s notable that, throughout my months of testing, every company emails me repeatedly – encouraging me to take further tests or update questionnaires about what I’m eating and how I’m exercising. It feels like a reminder that my data is of more value than my subscription fee.

Happily, the Thriva test doesn’t judge me at risk of either heart attack or diabetes. Unhappily, as with DNAFit, I’m not really sure I understand the results, despite the neat interface. Everything from liver function to lipid profile (which indicates your risk of developing heart disease) is presented on a gamified scale that goes from red to green. If a marker is on red, there’s an issue. If it’s green, as all mine are, then your results are within a healthy range. Again, there’s a lot of detail once you start reading. I’m not sure if I need to know that my cholesterol is 4.99 mmol/L. Or that my Total B12 is 273.0 pmol/L.

There’s no way to gamify giving a stool sample, but Atlas Biomed does make it as easy as possible by sending you a paper shelf to attach to the toilet to catch your deposit. It claims that it can map my microbiome – the ecosystem of bacteria, fungi, viruses and yeasts that live in my intestinal tract. Lately, the microbiome has been thought of as a key measure of internal health, analysing everything from our immune system to our stress and emotional responses.

Thriva told me what vitamins I may be deficient in; Atlas Biomed promises to help me “optimise” my ability to “synthesise” those vitamins. A few weeks later I get an email to say my results are ready. By this point I’m fairly well-versed in what to expect, but I can’t quite get over the fact that there is an online interface dedicated to my poo. And what a poo! Atlas Biomed even give it a nationality (Danish, because of my “high intake of flour-based foods, meat and fish, which is common throughout Europe”) and a character (village peasant – because I have a moderate amount of diversity in my gut bacteria, but not as much as I may have).

Unlike the other tests, everything does not look fine. I click through to the nutrition tab and find that I have a below-average microbiome ability to take in vitamins B1, B2, B3, B5, B6 and K. When I click on each vitamin I get an explanation of its function, and what could happen if I become very deficient. It tells me, for instance, that “vitamin B2 (riboflavin) participates in the formation of red blood cells and is also important for immune and reproductive health. Symptoms of riboflavin deficiency include painful cracked lips, mouth sores, swollen tongue, cracks in the corners of the mouth, skin lesions and muscle weakness. It may be accompanied by elevated homocysteine levels that can damage blood vessel walls, thus increasing the risk of cardiovascular disease and Alzheimer’s disease.”

Skin lesions and Alzheimer’s disease! The more I read, the more worried I am. I call my mother to tell her that it looks as though I’m going to suffer a slow and painful death.

“Are you ill?” she asks.

“No, but I’m not... well.”

She pauses. “You’re becoming very self-obsessed.”

***

My final test is the one I’ve been most curious about and the results are delivered in an almost old-fashioned way – by phone. “It’s the nurse from LetsGetChecked with your fertility results,” says a silky-voiced woman. I immediately worry. I never thought about children, really, until I got to 27. I wrote a story on the world’s largest sperm bank and the man who founded it told me that I was already old in childbearing terms. “By 30, all you’ve got left is the old rotten eggs,” he said. Charmer.

The LetsGetChecked site claims that testing the anti-Müllerian hormone (AMH) will give an indication of fertility by measuring how many eggs you have left. It’s another finger-prick test – and I dutifully fill up another vial of blood.

But a call feels like bad news. I sit down heavily. “Is everything…?” The nurse rushes to reassure me that everything is perfectly fine. “For AMH, we give you a score. Anything between 6.4 and 70.3 is considered normal,” she tells me. “Your score is 31.9.” I wait a beat. “And that’s good?” She explains that, for someone of my age, it is very good. I feel buoyed, and relieved that I’ve got a bit more mileage, even though I hadn’t given much thought to what I would have done if the test had shown I was low on eggs. I had just started dating someone; how practical is it to have a child with someone you’ve known only six months?

It turns out, however, that the science around AMH is in question, too. Last year the Journal of the American Medical Association published a study that found AMH levels had little bearing on a woman’s likelihood of conceiving within a year. As the researchers wrote, there “may be little association between a woman’s ovarian reserve and factors affecting fertility, such as egg quality”.

Finally, I decide it’s time to face up to what might be the most important test, the 23andMe test for the BRCA1/BRCA2 gene. I hover over the button for half an hour before finally clicking, and let out an audible sigh of relief when it tells me I do not have them. I feel lucky, as if I’ve won some kind of genetic lottery, and can’t help but relax about my occasional cigarette or night of drinking. But then I dig further; according to Harvard University researchers, 23andMe’s test is incredibly limited in scope, analysing three mutations across the two genes. But there are nearly 200 other mutations, which appear in dozens of other genes, which have also been linked to altered breast cancer risk. What’s more, just 5-10% of breast cancers are considered to be passed on genetically; in the other 90-95% of cases, environmental factors are the cause. Either way, I’m not off the hook.

After almost six months of testing, I am both wiser and more confused about my health. What did I learn? I’ve expanded my diet to include more leafy vegetables, and switched my workout regime to include longer runs and lower weights – and feel brighter and fitter for it. But while I have a lot more information, I’m not sure what to do with it. Even if the technology is cutting-edge, the advice remains broadly the same: there’s either nothing I can do, or I should live a healthier life. And I don’t need an online dashboard – even a very nice, primary-coloured one – to teach me that.

Which test does what?

23andMe genetics
Cost £149 (for health and ancestry)
Type of test Saliva
What it promises “90+ personalised reports about your ancestry, health, wellness and more.”
What it delivers The most comprehensive results, but most feel arbitrary – why would I need to know what colour my eyes are most likely to be? You’re unlikely to have heard of most of the genetic conditions, so it’s hard to see how this could be useful to many.

DNAFit genetics
Cost £179
Type of test Cheek swab
What it promises “Discover how your genes affect your fitness and nutrition... unlock your full health potential.”
What it delivers Data that, on first reading, doesn’t seem to have much real-world application. It includes a very good consultation with a sports or nutrition professional, though, who can help you write a personalised plan.

Thriva
Cost £69 for an advanced test
Type of test Blood sample
What it promises The advanced test checks for “signs of heart disease, liver disease and diabetes. Plus a look at four key nutrients required for optimal health.”
What it delivers A quick and sleek service, with green for “good”, red for “bad”, plus personalised advice from a GP and tips on what action to take (some of which are a bit patronising).

Atlas Biomed microbiome
Cost £139
Type of test Stool sample
What it promises “Learn how microbes can protect your gut from disease; understand how your diet affects your gut bacteria.”
What it delivers At the very least, you’ll come away with a better understanding of what various vitamins do. Plus the food recommendations are updated weekly and suggest interesting new foods to try.

LetsGetChecked ovarian reserve test
Cost £119
Type of test Blood sample
What it promises “An insight into how many eggs you have left in your ovaries by testing the anti-Müllerian hormone.”
What it delivers A dashboard that gives you the figure. Also a phone call from a very nice nurse who’ll explain your results before you get them, making everything easier to understand.

• Commenting on this piece? If you would like your comment to be considered for inclusion on Weekend magazine’s letters page in print, please email weekend@theguardian.com, including your name and address (not for publication).

• This piece was amended on 22 September to correct the costs of the Thriva and DNAFit tests