Highlights:
In 2018 scientists from Gladyshev lab specializing in
aging research started a collaboration with Insilico Medicine resulting
in a widely-publicized proof of concept microbiomic aging clock ;
The clock has been validated on multiple independent data sets;
The clock was shown to be biologically-relevant and used to
demonstrate that diabetic patients look older than their chronological
age;
The study was published in iScience and is expected to be used in new data analysis tools for COVID-19 and longevity research;
June 11th, 2020 - Recent advances in deep learning have allowed AI
algorithms to outperform humans in image, text, and voice recognition.
One particular use for AI in biology is deep aging clocks. Deep aging
clocks are trained on large samples to predict human biological age
using different data types, such as: pictures, videos, voice, blood
biochemistry, gene and protein expression, and MRI. In a
study recently published in
iScience,
Harvard and
Insilico Medicine
scientists used thousands of whole genome sequencing samples from gut
bacteria to develop and validate a new deep microbiomic aging clock.
This new tool indicates that the age of the host is a significant
contributor to the gut community dynamics.
Over the last decade human gut microbiome studies have produced
multiple surprising results. The bacteria in our gut are now known to be
major contributors to the immune function, brain development and
activity,
central metabolism, obesity pathogenesis and many other processes. The
growing realization of the role microbiota plays in human health makes
it essential to understand what factors shape gut communities and how to
manipulate them.
Such factors include the mode of birth, diet, physical activity
and age. The effect age elicits on microflora dynamics is much better
understood for the early stages of life. During the first year of life
all people are much more similar in terms of diet and behaviour,
compared to adults. Consequently, their gut flora goes through clearly
defined stages. But upon transitioning to adulthood, multiple
confounders such as diet, tobacco and alcohol consumption, and level of
physical activity make individual microfloras extremely diverse. The
NIH Human Microbiome Project
has shown that there is no core community in adult guts, although the
various combinations of microbial species tend to have similar functions
and metabolic capabilities.
Multiple studies have identified some age-related trends in gut
microflora. However, the findings usually have unclear general
applicability due to localized sampling. In a joint project between
Insilico Medicine and the laboratory of Vadim Gladyshev at Brigham and
Women's Hospital and Harvard Medical School, the data from 13 public
studies on human gut microbiome were aggregated to explore the
possibility of developing an aging clock based on the microflora
relative abundance profiles.
The initial attempt to predict chronological age based on gut
community species composition was published in BioRxiv in December 2018.
Since then the team further improved their approach and recently
published their results in the iScience journal. More than 1100
species-level microflora compositions were used to train a Deep Neural
Network in a cross-validated manner. The resulting ensemble predicts
hosts' age in an independent data set collection with a mean error of
5.9-6.8 years.
The published intestinal age predictor proves that there are
microflora succession patterns associated with age progression in the
adult. The described workflow can be used to recreate similar models
with data from other platforms and explore the effect of specific
bacterial taxa on the course of human aging in a more controlled
setting. The authors also suggest the ways to identify the microbes with
potential to accelerate or slow down aging.
"We are happy to collaborate with the Gladyshev lab on
this new microbiomic aging clock, which is the first of its kind. The
development of this clock was a long and tedious journey as we
originally thought that it would be impossible to build and after the
demonstration of the first proof of concept, it took two years to refine
and validate. We hope that the demonstrated approach will be used for
COVID-19 research and later for longevity research for tracking the
effects of different interventions and foods on the predicted intestinal
age", said Alex Zhavoronkov, PhD, CEO of Insilico Medicine.
The reported aging clock can be accessed at aging.AI. Insilico
Medicine aims to continue developing microbiomic tools and is planning
to release COVIDOMIC -- a tool for exploring variables with an effect on
the COVID-19 infection outcome, including those derived from patients'
respiratory microbiome.
###
Read the original research paper here:
https://www.sciencedirect.com/science/article/pii/S2589004220303849
About Insilico Medicine
Since 2014 Insilico Medicine is focusing on generative models,
reinforcement learning (RL), and other modern machine learning
techniques for the generation of new molecular structures with the
specified parameters, generation of synthetic biological data, target
identification, and prediction of clinical trials outcomes. Recently,
Insilico Medicine
secured $37 million
in series B funding. Since its inception, Insilico Medicine raised over
$52 million, published over 80 peer-reviewed papers, applied for over
25 patents, and received multiple industry awards. Website
http://insilico.com/.
For further information, images or interviews, please contact:
ai@insilico.com.
