Skip to main content
Premium Trial:

Request an Annual Quote

Science Studies Describe Machine Learning Algorithm to Speed Mendelian Diagnoses, Safe Drug Response Gene, More

A team led by researchers from Stanford University have developed a machine learning algorithm that can rapidly parse medical literature to speed the diagnosis of Mendelian disorders. As reported in Science Translational Medicine this week, the algorithm — called AMELIE, short for Automatic Mendelian Literature Evaluation — parses all 29 million PubMed abstracts, as well as more than 500,000 full-text articles, to identify information supporting the causality and associated phenotypes of most published genetic variants. It then prioritizes patient candidate variants for their likelihood of explaining any patient's given set of phenotypes. When applied to 215 diagnosed singleton Mendelian patients from the Deciphering Developmental Disorders project, AMELIE identified the causative gene in 66 percent of the patients. Its developers say their tool is three- to 19-times more efficient than hand-curated database-based approaches and offer it online.

A Washington University School of Medicine-led team has identified a gene required for a patient's safe response to nitrogen-containing bisphosphonates (N-BPs), a class of medications for bone disorders such as osteoporosis. In a study appearing in Science Translational Medicine, the researchers note that, while largely effective, N-BPs such as alendronate carry the risk of rare but traumatic side effects such as atypical femoral fracture (AFF) and osteonecrosis of the jaw (ONJ). Using CRISPRi-based genome-wide screening in cells and patients, they identify a gene called ATRAID that is required for alendronate inhibition of osteoclast function and, when knocked out in mice models of osteoporosis, impairs therapeutic responses to the drug. By sequencing the exomes of patients taking N-BPs who experienced AFF or ONJ, the investigators link differential expression of ATRAID to the side effects and poorer treatment outcomes. 

By studying the genomics of different goat populations, scientists from Northwest A&F University in China have uncovered new details about the animal's domestication. The researchers sequenced and analyzed the genomes of 164 modern domestic goats, 52 ancient goats, 24 modern wild goats, and four ancient wild goats from different locations worldwide. Among their findings is evidence of an ancient introgression event from a West Caucasian tur-like species to the ancestor of domestic goats. One introgressed locus with a strong signature of selection harbors a gene encoding a gastrointestinally secreted mucin that forms a protective glycoprotein coat involved in host innate immune responses to the invasion of multiple gastrointestinal pathogens.