Microproteins: How DNA’s dark matter is reshaping the drugs of the future
Giovanni Colonna , Domenico Aprile
AIMS Biophysics ›› 2026, Vol. 13 ›› Issue (2) : 225 -230.
Approximately 1000 to 4000 genes that encode small proteins (less than 100 amino acids) are still officially uncataloged in the human proteome. These sequences, often called smORFs (small open reading frames), have eluded identification for decades, which has relegated them to the simple category of “noncoding RNA” [1]. These sequences have eluded identification for three main reasons: 1. Arbitrary criteria: Historically, bioinformaticians have set a minimum threshold of 100 amino acids to exclude genetic “noise”. 2. Technical difficulties: Small proteins degrade rapidly and are difficult to detect using standard mass spectrometry. 3. Misclassification: Many are hidden within regions previously classified as “noncoding RNA” (ncRNA) or in segments overlapping with other genes. Today, we know that this “genetic noise” hides precision regulators critical to molecular biology, development, and metabolism [2]. Integrating polymer physics (physical-chemical properties) and deep learning (AlphaFold) is one of the few ways to scale up the study of the 1000−4000 missing microproteins. Without these models, we would have to study them one by one with crystallography, which would take centuries.
| [1] |
|
| [2] |
|
| [3] |
|
| [4] |
|
| [5] |
|
| [6] |
|
| [7] |
|
| [8] |
|
| [9] |
|
| [10] |
|
/
| 〈 |
|
〉 |
PDF
(368KB)
