Cell Culture Shots briefings – deep learning applications

machine learning algorithms - predicting biological age and protein structures

Immunological clock

People with a better functioning immune system can live longer. Scientists have long been trying to explore the markers that predict a person’s health or biological age. Measuring epigenetic changes on DNA is one known strategy but complicated. How about measuring the markers from the blood test – isn’t it simple.

As we age, the cells become damaged and emit inflammation-causing molecules. The healthy individuals with well functioning immune system can neutralize, while the other will age faster. The scientists analyzed blood samples from 1001 subjects of various ages and studied the chronic systemic inflammation to understand the markers. Using machine learning by feeding the protein biomarkers, they were able to identify CXCL9 as a top contributor.

The scientists grew endothelial cells in the lab and made them aged faster by repeated divisions to validate. High expression levels of CXCL9 from aged cells made the cells dysfunctional; however, the cells regained the function when silenced.

The inflammatory aging clock (iAge) functions through machine learning is the first tool to assess health by measuring the inflammatory markers.

Explore more by reading the  full article

Open science is the need of the day. Specifically, when two rival companies working on significant aspects of answering critical problems and sharing the code with the public is a big step in science. Compare mobile phone features vs. cost five years before and now. The features we see are a massive advantage to the customers.

The same is true when two academic groups shared the source code, which predicts the 3D shape of the protein accurately. This significantly reduces the time needed to crystalize the proteins and cost associated with X-ray crystallography and cryo-electron microscopy. Besides, not all proteins are amenable to such analysis.

DeepMind made the breakthrough by introducing the AlphaFold 2 to accurately predict protein’s shape with just the amino acid sequence. The details of this tool was briefly presented to the CASP community (Critical Assessment of protein Structure Prediction). It has spurred new ideas to Baker’s and their colleagues, who recently launched RoseTTaFold.

Both the tools take few minutes to hours to accurately predict the structures of proteins, which used to take some days.  DeepMind collaborated with several researchers working on neglected diseases. Good days ahead for structural biologists and drug discovery researchers – working on proteins that do not have a crystal structure.

Read the details published by two groups:

Baek, M. et al. Science https://doi.org/10.1126/science.abj8754 (2021).

Jumper, J. et al. Nature https://doi.org/10.1038/s41586-021-03819-2 (2021)

While we have machine learning tools helping us determine the age and protein structure, we are still searching for alternate animal models to test the drugs. The US government is investing money to breed monkeys for biomedical research. Mainly during COVID-19, scores of vaccines were tested in monkeys before being given to humans. 

Rhesus macaques (Macaca mulatta) are the standard non-human primate models used to study infectious diseases. Since they show genetic and physiologic resemblance to humans. On one side, animal-right groups seeking to stop the usage of animals in research – working against the scientists who insist that animal models are inevitable to understand the scores of human disease conditions.

The future is for organoids and deep learning algorithms that help reduce the animal models in predicting the efficacy and toxicity of the new therapeutic candidates. 

The mind and consciousness are not just in the head; rather, it is an essential element in every cell of a human being. It is often inquired and discussed a lot by both yogic and neuroscience scientists. 

It looks like the essential defense function is also attributed to each cell to a certain extent – rather than just the immune cells. Yep, this may be extrapolation, but there are many proteomes that we need to understand their function. The function of many proteins is not very clear. 

Apolipoproteins are secreted proteins found in extracellular spaces and known to transport cholesterol and lipids throughout the body. The presence of few family members of apolipoproteins in mammalian cells is not explained much. Scientists investigated and found that APOL3 protein can fight bacteria that enter the cells. APOL3 can bind to the lipids in the inner bacterial membrane and break apart, destroying the cells. 

Previously, the other family member of apolipoprotein APOL1 shown similar activity against the trypanosomes. The investigators are trying to understand the range of antimicrobial activity of these proteins – covering antiviral and antiparasitic studies.

We may find many unknown functions of the protein we know and many new proteins with new functions in the future.