Category Biology/Biotechnology

Mutations to Dyrk1a gene lead to brain undergrowth; an existing drug rescues the condition in newborn mice. Damage to the autism-associated gene Dyrk1a, sets off a cascade of problems in developing mouse brains, resulting in abnormal growth-factor signaling, undergrowth of neurons, smaller-than-average brain size, and, eventually, autism-like behaviors, a new study from Scripps Research, Florida, finds.

The study from neuroscientist Damon Page, PhD, describes a new mechanis...

Enzyme with central role in cancer and type 2 diabetes also activates. Researchers have discovered a direct link between a master sensor of cell stress and a protein that protects the power stations of cells. The same pathway is also tied to type 2 diabetes and cancer, which could open a new avenue for treating all three diseases.

When cells are stressed, chemical alarms go off, setting in motion a flurry of activity that protects the cell’s most important players. During the rush, a protein called Parkin hurries to protect the mitochondria, the power stations that generate energy for the cell. Now Salk researchers have discovered a direct link between a master sensor of cell stress and Parkin itself...

Scientists detail new work on NLRP3, an intracellular complex that has been found to participate in melanoma-mediated inflammation, leading to tumor growth and progression. By inhibiting NLRP3, the researchers found, they can reduce inflammation and the resultant tumor expansion.

Cancer doctors may soon have a new tool for treating melanoma and other types of cancer, thanks to work being done by researchers at the University of Colorado Cancer Center.

Specifically, NLRP3 promotes inflammation by inducing the maturation and release of interleukin-1-beta, a cytokine that causes inflammation as part of the normal immune response to infection. In cancer, however, inflammation can cause tumors to grow and spread.

“NLRP3 is a member of a larger family that is involved in sensing da...

Rice University computer scientists have demonstrated artificial intelligence (AI) software that runs on commodity processors and trains deep neural networks 15 times faster than platforms based on graphics processors.

“The cost of training is the actual bottleneck in AI,” said Anshumali Shrivastava, an assistant professor of computer science at Rice’s Brown School of Engineering. “Companies are spending millions of dollars a week just to train and fine-tune their AI workloads.”

Shrivastava and collaborators from Rice and Intel will present research that addresses that bottleneck April 8 at the machine learning systems conference MLSys.

Deep neural networ...