Scientists know little about specific hereditary factors leading to amyotrophic lateral sclerosis (ALS). However, a new artificial intelligence (AI) project from Germany may help unravel more answers.
A group of researchers from Bielefeld University has developed DiseaseCapsule to help learn more about the genetic factors that contribute to ALS.
Their findings were published on February 13 in Nature Machine Intelligence.
ALS is commonly referred to as Lou Gehrig’s Disease in honor of the former Major League Baseball star in the 1930s who was forced to retire due to the illness. The disease affects the nerve cells which help muscles operate in the lower and upper regions of the body, making ALS a type of motor neuron disease.
Early symptoms of the disease include muscle twitches, muscle cramps, muscle stiffness, altered speech, and difficulty chewing. As ALS progresses, individuals may find early onset signs worsen along with difficulty getting out of bed.
The National Institute of Neurological Disease and Stroke says that ALS may sometimes lead to language trouble, decision-making, and even dementia.
Registry data from the CDC regarding ongoing ALS cases shows a potential of 17,800 to 31,843. Currently, ALS is not a reportable disease in the majority of states within the U.S., nor is the CDC notified of new cases.
Typically, individuals become aware of ALS between ages 55 and 75. The CDC says nearly 5% to 10% of ALS cases occur within families, known as familial ALS. Nearly 25% to 40% of all cases of familial ALS are caused by a defect in the C9ORF72 gene, a vital protein found in motor neurons and nerve cells in the brain. No gender is more impacted by ALS, as it can be found equally in both men and women.
Although there is no cure for ALS, researchers from Bielefeld University highlight ways to improve the life quality of ALS patients. One study mentioned in their investigation shows nicotinamide adenine dinucleotide (NAD+) supplements may provide improvement to ALS patients.
A major conundrum for ALS patients is the lack of early detection, leading to missed opportunities for early treatment. But If ALS can be detected earlier on, perhaps rising treatments may be able to benefit ALS patients. To become possible, hereditary factors leading to the disease must be discovered.
Many hereditary factors contributing to ALS are unknown, giving reason to the new project led by Alexander Schönhuth, M.D., of Bielefeld University’s Faculty of Technology. In a university press release, Schönhut says 80% of ALS’s heritability' is still unexplained.
Schönhut and his team used an artificial intelligence method known as capsule networks (CapsNets) to analyze the genetic information of 3,000 patients with ALS and 7,000 as control subjects without the disease.
DiseaseCapsul is the first of its kind to operate across the whole genome and not focus on a few select core disease-related genes. Also, DiseaseCapsul is efficient in revealing the hierarchical structures of genetic interactions.
"The great advantage of this method is that it can capture overlapping processes," Schönhut said. "Our AI method, in contrast, clearly and comprehensibly shows which genes and their processes are particularly important for the development of ALS."
Numbers show the DiseaseCapsule to predict with 87% accuracy whether an individual will develop ALS. This is a 28% improvement over PRS (Polygenic Risk Score), the current clinical standard for predicting ALS.
Researchers were successful in discovering 922 candidate genes for association with ALS. Many of the candidate genes uncovered had not been unveiled by following conventional genome-wide association studies protocols.
Despite this, DiseaseCapsule identifies only 64% of the ALS patients currently undiscovered by PRS. The disease capsule excels at uncovering ALS but may not be as effective in benefiting those needing early intervention.
Although the device is not perfect, Schönhut anticipates ALS patients will soon benefit from the DiseaseCapsul technology.
"Each gene is engaged in different biological processes: the more we learn about the genes, the more we learn about the processes," Schönhut said. "In this way, our results will help people affected by ALS to adapt their lifestyle and reduce their risk of suffering from the disease. In addition, drugs could also be developed that influence specific processes."
- Nature Machine Intelligence. Predicting the prevalence of complex genetic diseases from individual genotype profiles using capsule networks
- CDC. What is Amyotrophic lateral sclerosis (ALS)?
- Science Direct. NAD+ in Brain Aging and Neurodegenerative Disorders