Genetic activity in blood could help detect ALS, predict survival: Study
Novel approach uses machine learning to ID biomarkers, potential treatments
- New method uses blood genetic activity, machine learning to detect ALS, predict survival.
- A 46-gene panel accurately IDs ALS, showing promise for improving diagnosis.
- Analysis identifies potential ALS treatments, including existing drugs.
Researchers have developed a novel approach to detect amyotrophic lateral sclerosis (ALS) and predict survival outcomes in patients by measuring genetic activity in blood cells, according to a study.
The scientists also built on their genetic activity analysis to define biological pathways that are disrupted in ALS and identify potential treatments for the disease.
“Our findings present an incredible opportunity to potentially diagnose ALS earlier, which opens up doors to treatments and clinical trials for which people otherwise may not be eligible due to advanced disease,” Eva L. Feldman, MD, PhD, co-author of the study at the University of Michigan, said in a university news story.
The study, “Gene expression signatures from whole blood predict amyotrophic lateral sclerosis case status and survival,” was published in Nature Communications.
Researchers measured activity of genes that code for ALS biomarkers
ALS is characterized by the death and degeneration of motor neurons, the nerve cells that control movement, resulting in progressive muscle weakness.
Available treatments are limited, and most ALS patients only live a few years after disease onset. However, there are still no objective tests that can reliably diagnose ALS or predict how long people with the disease are likely to survive.
“In the current study, we sought to address these unmet needs by developing blood-based gene expression signatures of ALS risk and survival, along with therapeutic drug candidates using drug perturbation analysis,” the researchers wrote.
Normally, researchers measure the levels of specific disease biomarkers in the blood or other bodily fluids to determine if they can detect ALS. But here, the team instead assessed the activity of the genes that code for those biomarkers, measuring which genes were active and to what extent.
The researchers collected blood samples from 422 people with ALS and 272 controls without the disease and used an RNA sequencing method that detected the activity of more than 22,000 genes. They then used machine learning to analyze the differences.
Machine learning is a type of artificial intelligence in which a computer is fed a large data set and uses algorithms to identify patterns within the data. The resulting models can then be applied to make predictions on future datasets.
For this analysis, the machine learning models were trained using data from most of the ALS patients and controls, and then the remaining samples were used to test the models.
46-gene panel could ID ALS with 91% accuracy
After some refinement, the researchers ended up with a 46-gene panel that could identify ALS with roughly 91% accuracy in their data set. In an external data set, which lacked data on several of these genes, the researchers’ algorithm still performed reasonably well, with more than 60% accuracy.
“After testing our model on our own samples, as well as data from other groups, it performed better than any previous attempt at an ALS biomarker signature,” said Yue Zhao, PhD, first author and research assistant professor at the University of Michigan. “Our results suggest a need for further investigation into this model as a tool to improve diagnostic accuracy and decrease diagnostic delay.”
The researchers also conducted analyses in which they paired the genetic activity data with clinical and demographic data, such as age at symptom onset, the age at which muscle weakness first appeared, and sex, to predict survival outcomes. These models likewise showed high accuracy.
Pursuing these important next steps has incredible potential to advance diagnostic and therapeutic opportunities in ALS that could ultimately improve clinical care. It is an exciting time in ALS research.
Building on these findings, the researchers conducted additional analyses of the ALS genetic data to identify biological pathways that are dysregulated in the disease. They then extrapolated those findings to identify potential ALS treatments.
They identified eight potential treatments, including some medicines that are already in use: trifluoperazine, an antipsychotic used to treat schizophrenia, and ibrutinib, which is used to treat certain blood cancers.
The scientists said that further studies are necessary to confirm whether these medicines may be beneficial in ALS, but they noted that these analyses provide promising directions for future research.
“Pursuing these important next steps has incredible potential to advance diagnostic and therapeutic opportunities in ALS that could ultimately improve clinical care,” said Maureen A. Sartor, PhD, co-senior author of the study at Michigan. “It is an exciting time in ALS research.”
About the Author
