We have developed our epigenetic tests with leading research universities, such as Harvard, Yale, and Duke. The DunedinPACE algorithm, which we license from Duke, is widely considered the most precise and predictive aging measurement (epigenetic clock) in the world.

Epigenetic-based analytics is one of the biggest growing areas of functional medicine and has been written about in the best news outlets due to its potential (Nature, Forbes, The Guardian, and the Wall Street Journal, and USA Today).

We only offer & recommend clocks with the following criteria:

• Are 2nd or 3rd generation clocks: We only recommend clocks that are trained on biological phenotypes instead of chronological age. Chronological age clocks generally have very low predictive power to disease outcomes and therefore miss the largest point of these clocks which is to predict the biggest risk factors of all chronic disease and death. For instance, for one standard deviation, the DunedinPACE (3rd generation) increase risk of death by 62% versus the original horvath which increases risk by 2%.
• Have been published: Anyone can create a clock but without validation, it is essentially like going to a fortune teller. You can simply choose to believe or not. Published clocks show validation to disease outcomes which we think is imperative to show why it is important. This also impacts consumers, if the results don’t track with known lifestyle improvements and changes, you will have a lot of angry customers.
• Have high intraclass correlations for high precision: Some of the original clocks could vary by 40% on the same sample. This means, if you are 50 and take a test, then take the exact same sample and test again; you could have differences of 20 years in age. This can make customers extremely angry and makes it impossible to track change. With longitudinal testing being a feature of the testing this is a huge issue. This has long been an issue with the clocks as described in this paper. Now, all of our clocks have less than a 4% difference when testing the same sample multiple times.
• Have high hazard ratios to disease and healthspan: The best clocks are the most predictive, our clocks are more predictive of outcomes than any other clocks.
• Use whole blood: While people would love to do this in Saliva, it is not really possible. This is for multiple reasons. Generally, the biggest reason goes to validation. Most biobanks don’t bank saliva samples like they do blood. So, if you create an algorithm in saliva, it makes it very hard to validate with HR to outcomes. The only biobank with saliva currently is the LBC. Unfortunately, they have 450k array data. This means that algorithms derived from 850k need imputation, and we already know from the array point above, this causes big issues in TRUE validation. You can see algorithms across tissue comparison in this publication (average differences of almost 30 years observed in some age clocks).
• Include immune deconvolution methods: Just as algorithms in saliva are not compatible with blood. Every cell has a different aging signature. If we have different cell types in blood from one algorithm to the next, it can change the result. We have developed 12 cell immune deconvolution algorithms which can factor this in for much more precise results. You can see this here.
• Have shown to change with classical aging interventions such as caloric restriction: If people change their lives in a positive way but have their ages go up, it is the easiest way to lose a user. Thus, we need to show that these clocks are responsive and responsive to the things we already know impact lifespan. **In this paper with Yale,** we looked at 51 intervention trials and showed that Symphony and DunedinPACE are the most responsive to change. However, we have also shown that 1st generation clocks go in the WRONG direction with caloric restriction which is the most well validated interventional therapy. This is another reason NOT to use 1st generation clocks. Also, GrimAge was not significantly changed in the CALERIE trial.
• Show “How you are aging” with GenX (Generation Examplainable clocks): With previous clocks, you could tell if someone was older or younger. However, you could not tell them why. Now, we have generation explainable clocks. These clocks can explain why your aging has increased and make recommendations to improve. This is important because not everyone ages in the same way. Smokers might have higher lung, heart, and inflammatory ages while someone overweight might have higher hormone and metabolic ages. Each requires very different interventions to improve!
• Works on MSA and EPIC arrays: We also would like to stress that array type is very important here as well. EPICv1 trained algorithms like DunedinPACE, OMICmAge, SymphonyAge, GrimAge, and most others, do not have probes that are needed on the EPICv2. The only way to solve this is through imputation or adding probes to the array. We use a custom EPICv2 array to add the probes, as we don’t believe the imputation of the missing CpGs is the correct method. The problem with array compatibility has also been published by two different groups in studies here and here. Our algorithms work on every array type.

Generally, there are only 4 clocks that hit this criteria. These are SymphonyAge (developed with Yale), OMICmAge (Developed with Harvard), and DunedinPACE (Developed with Duke).  We also have GrimAgev2 which is the only one we don’t have exclusive commercial rights to. It is also very good.

I have also included the comparison between our algorithms and testing versus every other commercial testing company in the presentation below! We are the only company with published aging algorithms, only 2nd or 3rd generation clocks, our own CLIA lab, and quite a bit more. I hope this helps a little further.