Using data from hundreds of cancer patients, a team from Leiden University Medical Center has developed a neural network that can predict how an individual will respond to commonly used chemotherapeutics. The enzyme CYP2D6 is involved in the metabolism of around 30 percent of commonly prescribed drugs but its activity varies between individuals, impacting how certain drugs will affect them. While clinicians use genetic biomarkers to determine a patient's CYP2D6 profile, the approach is imperfect. As an alternative, the researchers trained a neural network on full CYP2D6 gene sequences from 561 breast cancer patients to model enzyme activity on a continuous scale. As reported in Science Translational Medicine this week, the network was able to predict individual drug response better than genetic profiling in an independent cohort of patients treated with tamoxifen or venlafaxine, both of which are CYP2D6 substrates. "These results demonstrate the advantage of a continuous scale and a completely phased genotype for prediction of CYP2D6 enzyme activity and could potentially enable more accurate prediction of individual drug response," the researchers write.
Increased fluctuations in gene expression, also known as noise, can trigger cellular reprogramming in mouse embryonic stem cells, according to a new study in this week's Science. While noise is often considered detrimental, it can also be used for the benefit of an organism. In the study, a University of California, San Francisco-led team performed a series of screens that revealed a compound called IdU that increased gene expression noise in a range of different cells. In mouse embryonic stem cells, specifically, IdU amplified the noise of gene expression without changing the overall rate of transcription of most genes through its interaction with the DNA-repair protein Apex1. Additional experimentation showed that increasing transcriptional noise with IdU increased cellular plasticity, leading to a higher frequency of embryonic cell reprogramming. "The ability to independently control the mean and variance of gene expression my indicate that cells can amplify transcriptional noise for fate exploration and specification," the study's authors write.