In an earlier video, I presented data for total cholesterol (TC) levels in blood in terms of changes during aging and all-cause mortality risk. I’ve measured TC 25 times in the past 5 years, and in this video, I present that data, and my approach to optimize it.
https://www.youtube.com/watch?v=PBv_hXwUqHM&feature=emb_logo
Michael Lustgarten 
Thank you for yet another very insightful post. Sorry … tons of questions/comments…
Why have you decided to look at the correlation between TC and Glucose and Creatinine? (i) Were those correlations the highest you have in your TC data or (ii) the ones you have ACM data for or (iii) quite rightly you have chosen between serious morbidities such as CVD, diabetes and kidney disease?
What if you had found a low correlation? Would you have chosen to yet optimize your TC vs ACM by increasing TC (the “battle 1”)?
Looking at ACM is important. But I wonder about specific genetic risk factors: e.g. reducing saturated FFA is, for specific genotypes, very effective to regulate problems with lipids. Have you looked at this too? This is where personalized nutrition would really matter. And, increasingly important, would be to consider the interaction between the host metabolism and microbiome.
Finally, I think AI tools should matter much more in guiding us finding unbiased correlations between a large number of variables, including the microbiome. Any thought on this?
Just out of curiosity: do you use a simple excel or a more sophisticated math pack to analyze your data?
Thank you again and good luck in your interesting journey.
FFA=free fatty acids (I measure in plasma regularly), TC=total cholesterol, ACM=all causes mortality, CVD=cardiovascular disease, AI=Artificial Intelligence
LikeLike
Ha, no need to apologize for comments and questions!
Why have you decided to look at the correlation between TC and Glucose and Creatinine?
-I’m always looking for correlations, whether it’s between blood biomarkers with diet, or with other blood biomarkers, to try to elucidate as much of the whole picture as possible. There was also a strong correlation between RBCs with TC, but the combination of glucose+creatinine’s correlation was stronger, so I included that in the story. I also tried to include RBCs with glucose and creatinine in a multivariate model, but when included, model significance was weaker than the combination of glucose + creatinine.
In terms of your genetics question, I’m not recommending everyone reduce saturated fat, but instead, to blood test themselves regularly and track diet, to try to discern the optimal diet for them. More to your point, I eagerly look forward to the day when an AI or ML-based approach will use data for my genome, dietary intake, circulating, urinary, and fecal metabolome, fecal microbiome, physical activity levels, etc. to derive the diet, supplement, exercise approach that will be predicted to maximize my lifespan. For now, I’m making educated guesses based on my own data to get close to that.
I’m a huge proponent using AI or ML in this process, and I recognize that I’m using a rudimentary statistical approach relative to that. At some point, I may learn to code to that I can apply AI or ML to my own data…
Simple Excel sheet for now, and thanks albedo!
LikeLike
Thank you! I am happy to join the similar AI/ML/DL journey too to same scope. I think you are making highly educated guesses. I have maybe a less rigorous approach than yours with data over ~20 years. Good luck!
LikeLike