AI Analysis Finds Serious Covid-19 Genetic Driver

(The Digital Artist, Pixabay)

27 Oct. 2021. A deep and multi-variable analysis of patients with serious Covid-19 respiratory symptoms finds the condition associated with abnormal expression of one gene. A team from the biotechnology company HiberCell Inc. in New York and University of Strasbourg in France report their findings in yesterday’s issue of the journal Science Translational Medicine.

Researchers led by Thomas Chittenden, HiberCell’s chief technology officer, and immunologist Seiamak Bahram at University of Strasbourg, are seeking to identify genetic drivers of serious Covid-19 disease, an elusive question due to the advanced age and comorbidities in many patients with the condition. In their study, the team investigated multiple molecular indicators in three sets of patients, all under 50 years of age: 47 patients in critical condition with acute respiratory distress syndrome, or ARDS, a complication of Covid-19 infections; 25 patients hospitalized with Covid-19 infections, but not in critical condition; and 22 healthy persons.

The team took blood samples from these individuals and collected data from the samples with a variety of techniques, including whole genome sequencing, RNA sequencing of whole blood, and protein analysis of plasma and white blood cells with nuclei, such as T, B. and natural killer cells in the immune system. The researchers also profiled cytokine enzymes produced by the immune systems of participants, as well as immune-system traits.

Algorithms, quantum computing, mathematical modeling

Data collected from participants were evaluated further with a number of computational techniques, including those drawn from artificial intelligence. Researchers used the data to train machine learning algorithms, supplemented with deep learning analytics to detect patterns in the data possibly missed by initial computations. The investigators also used quantum annealing techniques that use quantum computing to analyze complex multi-dimensional spaces for finding optimum solutions, and mathematical modeling to identify structural causes.

The analysis identified a network of some 600 genes affecting outcomes in ARDS patients, but further investigation highlighted abnormal over-expression of a single gene, called a disintegrin and a metalloprotease 9 or ADAM9, as the key driver of serious ARDS cases. ADAM9 is expressed in proteins that bind to cell membranes and are associated with inflammatory and degenerative diseases, as well as some tumors. The team verified its ADAM9 findings in separate groups of 81 critical-case patients with ARDS and 73 recovered patients. Further lab tests show decreasing ADAM9 expression reduces SARS-CoV-2 viral uptake and replication in human lung cells.

“This work provides a deeper understanding of COVID-19 in a younger, healthier patient subset,” says co-lead investigator Bahram in a HiberCell statement, who adds, “We are actively working to further elucidate the importance of ADAM9 in virus uptake and replication in human lung cells. These data may provide an alternative therapeutic approach for Covid-19 or in broader terms, ARDS in general.”

HiberCell is a spin-off enterprise from Mount Sinai medical school in New York, whose main work is applying machine learning and other artificial intelligence tools to find underlying causes of cancer, particularly its spread and relapse, and therapies to address those causes. In February 2019, Science & Enterprise reported on the company’s founding and first venture round that raised $60.8 million.

More from Science & Enterprise:

• Small Biz Grant to ID Digital Biomarkers for Alzheimer’s
• Partnership Seeks Lower-Cost Protein Drugs
• Animal Genome Biotech Raises $9M in Seed Funds
• Small Biz Grant Funds Superbug Therapy Technology
• Biotechs Partner on Agitation Detection Technology

*     *     *