AI News, BOOK REVIEW: Cracking Open the Black Box of AI with Cell Biology

Cracking Open the Black Box of AI with Cell Biology

The deep neural networks that power today’s artificial intelligence systems work in mysterious ways.

They’re black boxes: A question goes in (“Is this a photo of a cat?” “What’s the best next move in this game of Go?” “Should this self-driving car accelerate at this yellow light?”), and an answer comes out the other side.

But a new study that mapped a neural network to the components within a simple yeast cell allowed researchers to watch the AI system at work.

The trainers feed in a dataset (millions of cat and dog photos, millions of Go moves, millions of driver actions and outcomes), and the system connects the neurons in the layers to make structured sequences of computations.

So his team mapped the layers of a neural network to the components of a yeast cell, starting with the most microscopic elements (the nucleotides that make up its DNA), moving upward to larger structures such as ribosomes (which take instructions from the DNA and make proteins), and finally to organelles like the mitochondrion and nucleus (which run the cell’s operations).

DCell allows researchers to change a cell’s DNA (its genetic code) and see how those changes ripple upward to change its biological processes, and subsequent to that, cell growth and reproduction.

Its training data set consisted of several million examples of genetic mutations in real yeast cells, paired with information about the results of those mutations.

“If you could construct a whole working model of a human cell and run simulations on it,” says Ideker, “that would utterly revolutionize precision medicine and drug development.” Cancer is the most obvious disease to study, because each cancer patient’s tumor cells containa unique mix of mutations.

Researchers need to gather enough information about human patients to form a training data set for a neural network—they’ll need millions of records that include both patients’ genetic profiles and their health outcomes.

Cataloging a cancer cell’s biological processes is tough because the mutations don’t only switch cellular functions on and off, they can also dial them up or down, and can act in concert in complicated ways.

How a Yeast Cell Helps Crack Open the “Black Box” Behind Artificial Intelligence

“But all of these systems are so-called ‘black boxes.’ They can be very predictive, but we don’t actually know all that much about how they work.” Ideker gives an example: machine learning systems can analyze the online behaviors of millions of people to flag an individual as a potential “terrorist” or “suicide risk.” “Yet we have no idea how the machine reached that conclusion,” he said.

Ideker’s research team recently developed what they call a “visible” neural network and used it to build DCell, a model of a functioning brewer’s yeast cell, commonly used as a model in basic research.

They trained DCell on several million genotypes and found that the virtual cell could simulate cellular growth nearly as accurately a real cell grown in a laboratory.

But if the researchers instead wired ribosomes to an unrelated process like apoptosis, a system cells use to commit suicide, DCell could no longer predict cell growth.

“We want one day to be able to input your specific cancer-related genetic mutations and get back a readout on how aggressive your cancer is, and the best therapeutic approach to prevent its growth and metastasis,” said Ideker, who is also founder of the UC San Diego Center for Computational Biology and Bioinformatics.

Virtual Cell Can Simulate Cellular Growth Using Machine Learning

Scientists have created a virtual yeast cell model that can learn from real-world behaviors, a key step in utilizing artificial intelligence in healthcare to diagnose diseases.

team of researchers from the University of California San Diego has developed what they called a “visible” neural network that enabled them to build DCell—a machine learning model of a functioning brewer’s yeast cell that is commonly used in basic research.

The research team is now generating some of the experimental data they would need to build a DCell system for cancer and determine how to best personalize the virtual cell approach for each individual patient’s unique biology.

“We want one day to be able to input your specific cancer-related genetic mutations and get back a readout on how aggressive your cancer is, and the best therapeutic approach to prevent its growth and metastasis,” Ideker said.

A Network of Conserved Synthetic Lethal Interactions for Exploration of Precision Cancer Therapy

(B) Percent of patients in TCGA harboring either a somatic mutation (n = 6,911) or homozygous deletion (n = 7,462) in any of the TSGs chosen for screening.

The incidence of both somatic mutation and homozygous deletion is higher for the TSGs with yeast orthologs included in this study relative to a random set of genes (inset).

(D) For each TSG (x axis), the plot shows the fraction of druggable genes screened for synthetic lethal interactions in prior studies in yeast ( Ryan et al., 2012 ) (y axis).

Using Molecular Networks to Seed Predictive Hierarchies of Cell Systems

Trey Ideker, UC San Diego Network Biology

Dr. Eric Verdin on Ketogenic Diet Longevity, Beta-Hydroxybutyrate, HDAC Inhibitors & NAD+

Eric M. Verdin, M.D. is the fifth president and chief executive officer of the Buck Institute for Research on Aging and is a professor of Medicine at UCSF.