The very rare condition that a family pushed science to solve

Hutchinson-Gilford progeria syndrome is extremely rare. Children are usually born looking healthy, then within the first two years grow slowly and develop signs that look like early ageing, such as hair loss and stiff joints. Most serious of all, their arteries harden, causing heart disease and strokes at a young age. Their intelligence is not affected.

When their son Sam was diagnosed in 1998, the doctors Leslie Gordon and Scott Berns found almost no research on the condition. In 1999 they started the Progeria Research Foundation, which collected cells and data from children around the world. In 2003 researchers working with the foundation found that progeria is caused by a change in the LMNA gene, which makes a faulty protein that damages the cell nucleus. The change is usually new in the child, not inherited from the parents.

In November 2020 the US approved lonafarnib, sold as Zokinvy, the first treatment for progeria. Studies showed it extended average survival. Sam Berns became widely known for his outlook on life and died in 2014, aged 17.

Rare diseases are easy to overlook: there are few patients, little money and scattered data. Here, a determined family built the registry and samples that researchers needed, and progress came faster than anyone expected.

The AI angle: AI models learn from common cases and can miss rare ones, because few examples appear in their data. For rare conditions, specialist doctors and patient groups know far more than any general chatbot.

The idea behind this caseLesson: How machines learn →

1. 1Remember that rare conditions are poorly represented in most data, including the data AI learns from.
2. 2Look for patient organisations and specialist centres when a condition is rare.
3. 3Support research registries if a rare condition affects your family and you are comfortable sharing.
4. 4If a child is not growing as expected, see a doctor, who can refer you to specialists if needed.