DNA sequencing vs sequencing human movement - lessons from genetics

Those of you who have been following our startup journey have probably heard us refer to what we’re building with our Large Movement Models as “sequencing human movement”. Let’s dive into this analogy.

‘Sequencing’ in the genomic sense means extracting DNA, the building blocks of organic life, and determining the order of its component parts, base pairings of four chemicals (paired as Adenine-Thymine and Guanine-Cytosine for those of you who remember your biology). While these building blocks are the same across all known cellular life forms, the way in which they are put together dictates the organism. If we just look at modern humans, our genome consists of ~3 billion base pairs. Across two individuals (who aren’t twins) the genetic individuality between them is ~0.1%. Scaling up from chemical building blocks, modern humans are also composed of the same anatomical building blocks. Although there are exceptions, for the most part each of us consists of a musculoskeletal system with two feet, two legs, a torso, two arms, associated joints, and a head, controlled by a central nervous system. We have evolved to move bipedally, getting from place to place through propulsion and braking on alternating legs via an oscillating stance and swing phase. How the different component parts work together, however, varies. Compared to the next person, you may swing your foot further out to the side while you’re walking, or bounce up and down more, or swing your shoulders more than your hips, or brake with your knee more extended, etc.. None of this variability is necessarily bad, it is getting you from point A to point B after all, it’s just your way of walking. In other words, while the overall pattern of walking is consistent across humans, the details, or “sequence” of each individual’s movement patterns and mechanics are unique, just like our genetic code.

We can extend this analogy to pathology. Sequencing our genetic code enables us to look for known mutations, or errors in the order that can have impacts on health. We now know that we accumulate mutations as we age, which can lead to disease or pathology. Our movement mechanics also change over time as our bodies become less elastic and degrade. As these “mutations” accumulate, they can result in pathology, for example hip pain or low back pain. Identifying these “mutations” can help us mitigate cascading pathologies (for example, when your knee hurts so you shift to put less weight on it, then your opposite hip starts hurting, then your back starts hurting, etc.) and start preventative treatment. Genetic testing allows us to identify genetic diseases like Huntington’s and intervene earlier. There is a vast realm of research and commercial work surrounding genetic testing in predictive and precision medicine, looking for patterns in the code that indicate risk level for different diseases. By sequencing our movement at the individual level, we gain a level of precision previously unheard of. Similar to the power of large genomic datasets, large, detailed movement data sets open up the ability to begin looking for patterns that are indicators of disease - movement-based biomarkers.

The human genome was sequenced relatively recently, in 2003 (over 20 years ago now, yes everyone is feeling old), and has already resulted in seismic shifts in the way we approach health and medicine. Movement has the power to be the next such shift, that’s why we’re sequencing the human movement fingerprint 🧠