Health Testing Reviews for Impute

£0.00 Converted from $0.00

At a Glance

Editor's Rating:
4 out of 5 stars
Customer Service:
4 out of 5 stars
Clarity of Results:
2.5 out of 5 stars
References Cited:
5 out of 5 stars
Value for Money:
5 out of 5 stars

Summary

For a pay-what-you-can service relying on donations, the Impute website was very generous, offering free genomic imputation and analysis tools. The results provided by their GWAS and UK-biobank calculators, in particular, were very comprehensive. Still, their website was not especially user-friendly, and I felt their results wouldn’t be easy for everyone to understand.

Full Review

Impute.me is a non-profit website run by independent academics, offering genetics analysis. The site was established in 2015, and is led by Lasse Folkersen, the lead scientist at Sankt Hans Hospital, Denmark. They aim to provide up-to-the-minute genetic analysis using the latest research, and to present technical information in a way that is s user-friendly as possible.

Product Expectations

At first glance, the Impute.me site appears pretty basic. Clicking on “Modules” in the navigation bar, I found a list of the types of tests they offered, including a ‘Rare Diseases’ test, a ‘Mutation senser’ and something called ‘Politics’, which turned out to be a “sanity check” looking at whether genetics could predict a person’s political opinions.

I thought I would try the ‘Athletic Performance’ test, the ‘UK-Biobank Calculator’, and the ‘GWAS Calculator’. The Athletic Performance test would report on the genetic variants to do with exercise and fitness. I thought this test would be a good place to start, since I expected to have a reasonably good idea of the accuracy of the results.

I read that the two calculator tests would calculate a “genetic risk score” that would show my likelihood of developing different diseases. The UK-Biobank Calculator used a study of around half a million UK residents, while the GWAS calculator was based on “thousands of studies”.

Ordering Experience

It wasn’t incredibly obvious how to upload my genetic data. But in the “Modules” drop-down menu I found an item labelled “ImputeMe (start here)”.

Here, I found I could upload my data without the need to create an account, though I did need to provide an email address so that my results could be emailed to me.

I read that I could upload my genetic measurements “such as those provided by 23andMe or ancestry.com”. They would then use imputation technology to fill in “millions of additional genetic variations that were not measured in the original data.” There was a link to a video on Kickstarter to explain this process, which would use “overall knowledge of human ethnicity and ancestry” to make educated guesses about the gaps in my genetic data.

I was surprised to find that once I uploaded my data, the imputation process would take “a few days”, after which I’d receive an email with a login ID, which would enable me to use Impute’s genetic analysis tools.

I uploaded my genetic data file, which I had already downloaded from 23andMe. It took about a minute for the file to upload. I entered my email, and left the box ticked to allow my data to be deleted after two weeks. There was a link to the “Terms of use”.

Clicking this took me to a very basic webpage with no navigation options. The Terms of Use were very short. Still, they seemed reasonable and transparent enough. I learnt that the site was non-profit and intended for educational and research purposes. They claimed no responsibility for any medical interpretations that users might make from their results.

I read that they took my anonymity “much more seriously than any other online genetics-service”, and would require only an email address to send me my results. I could choose to have my genetic data deleted two weeks after imputation, though it would be deleted anyway if the server was full.

They reserved the right to contact participants by email to ask if they would be interested in “follow-up discussions on academic research”, and claimed these emails were “entirely ignorable”. Well… fine.

Back on the upload page, I hit “Start imputation”, and a message popped up telling me that it would typically take between one and five days for my genomic data to be processed, depending on server queue. I would receive an email with my unique user ID and download instructions. I was immediately sent an email telling me my data was queued for imputation.

A couple of days later, I received an email telling me my imputed genome was ready.

The Results

The email contained a link to download a copy of my imputed genome, and also a login ID for Impute.me, which would allow me to use their genetic analysis tools. There was a link to donate to the site using PayPal.

Results Section: Athletic Performance

I went ahead and tried out the Athletic Performance test. I entered my login ID and hit “Run analysis”. The analysis was almost instant. Two tables were generated, the first showing three genetic variants linked to athletic performance, and which genotype I had (shown below).

Click to see larger
A chart showing three genetic variants related to athletic performance.

A chart showing three genetic variants related to athletic performance.

According to the chart, I had one copy of the T allele (letter) in the ACTN3 gene, and one copy of the C allele. In the information, I read that the C allele was associated with better sprint performance, and that this was linked to solid evidence. Since I had one copy of the C allele, I benefitted somewhat from this.

For the ACE gene, which I read could potentially distinguish between a predisposition for power sports versus endurance training, my genotype hadn’t been imputed, and there was a note saying that imputation often failed for this variant.

For AGT, I read that the C allele had been associated with power sports performance. This meant little for me, though, since my genotype here was A/G.

A second table showed genetic risk score for various sports injuries and dietary needs (shown below).

Click to see larger
My genetic risk scores for injuries and nutrients.

My genetic risk scores for injuries and nutrients.

In the information, I read that the risk scores calculated were not my own risk score, but were actually percentiles showing which “percent of people have a lower score”. This was meant to show how I related to other people in terms of risk, though I wasn’t entirely sure whether this meant the percentage showed how many people had a lower score than me, or just had a low score.

I decided that my “Level-score” of 76.9% for stress fracture probably meant that I was in the upper percentile in terms of risk, and so had a higher risk of stress fracture than 76.9% of people.

For the vitamins listed, it wasn’t clear whether a high risk meant a risk of deficiency of that vitamin, though that would be the obvious interpretation. And I didn’t know how the “Number of SNPs” listed for each item contributed to my risk.

Results Section: GWAS Calculator

Next, I decided to try the GWAS calculator. Again, I had only to enter my login ID, and then run the analysis.

This reported on all sorts of traits, and so I will only mention a few here. The first was severe acne, and looked at 13 genetic variants in order to calculate my risk. My genetic risk score was shown in a chart, relative to the average risk score among the population (shown below).

Click to see larger
My genetic risk score for severe acne.

My genetic risk score for severe acne.

Since I’ve never had severe acne, I assumed that my score – between 0 and -1 – meant that I had a decreased risk. Looking at the information, I found that my risk had been calculated as -0.72. They stressed that this meant my genetic risk score was “higher than 23% of the general population”. At first, I thought this meant my score was higher than average, but then I realised that it was, in fact, lower than 77% of the population.

I had a look at the table showing the relevant SNPs (Single Nucleotide Polymorphisms) or genetic variants (shown below).

Click to see larger
A table showing the SNPs associated with severe acne.

A table showing the SNPs associated with severe acne.

At first, I thought this table was more or less impenetrable, until I worked out what to look for.

For each SNP, there was a column with “Your Genotype” and a column showing the “Risk/non-risk Allele”. For the first SNP listed (rs330071), the risk allele was given as G, and the non-risk allele as A. I saw that my genotype was G/G, and so I had inherited both copies of the risk allele, which would have raised my overall risk score for acne. For other SNPs, I had – for the most part – inherited the non-risk alleles, which was why my overall risk score was lower than average.

This was not as straightforward as it might have been, and the tool was clearly intended to be used by someone with a strong knowledge of genetics. Most of the information on the table meant very little to me.

I had a look through some of the other traits listed in the GWAS calculator.

I was intrigued to see that “Anger” was among the list, though some of my SNPs associated with anger were “missing or discrepant” since they hadn’t been imputed, and didn’t form part of my original data. All the same, I’d been assigned an overall risk score of -1.7, meaning that I was more predisposed to anger than 4% of the population, though less disposed than 96%.

Another interesting – and slightly peculiar – trait they had isolated was a predisposition for gambling. To my surprise, I scored pretty highly for this: higher than 99% of the population, in fact! Still, I read that having a high genetic score for something that isn’t very heritable “may make little difference”. Looking at the SNPs listed, I found I had both copies of the risk alleles for four of the five SNPs associated with gambling, and one copy of the risk allele for the remaining fifth.

Most of the traits listed by the GWAS calculator were to do with disease risk. There were plenty of risk calculations for different cancers, as well as mental illnesses such as major depression and bipolar disorder. But alongside these were scores for perfectly innocuous traits, such as eye colour, or for predicting things like sleep depth or resting heart rate.

Results Section: UK-Biobank Calculator

I found the UK-Biobank calculator quite similar to the GWAS calculator, though the population reference they used was only people living in the United Kingdom.

I also found they measured slightly odder things, which I would not have thought had much genetic predisposition involved at all, like wrist fractures (which I apparently had a slightly higher than average risk for – better watch out).

There were also some weirdly specific items, like “Fed-up feelings” and “Had menopause” (which would be inevitable in women over a certain age, and highly unlikely in men). There was also genetics relating to where you lived. Apparently, I was more likely to live in an “Accessible Small Town” in Scotland than 33% of the population.

A lot of the traits listed were to do with eye problems, and various reasons I might have for wearing glasses or contact lenses. For instance, my risk of wearing glasses for “a ‘squint’ or ‘turn’ in an eye since childhood” was higher than 38% of the population, though many of my SNPs for this hadn’t been imputed.

As well as various genetic predispositions to diseases, I found a few personality traits. I learned that I had a low risk score for being a “Worrier” suffering from anxious feelings, scoring lower than 97% of the population. Conversely, I found I had a high score for “Risk taking”, scoring higher than 73% of the population. Since these seemed like two sides of a coin, I thought it was right that I should score very low for one, and highly for the other.

Summary

For a pay-what-you-can service relying on donations, the Impute website was very generous, offering free genomic imputation and analysis tools. The results provided by their GWAS and UK-biobank calculators, in particular, were very comprehensive. Still, their website was not especially user-friendly, and I felt their results wouldn’t be easy for everyone to understand.

See a description of this DNA test from Impute >