CREME is a groundbreaking virtual laboratory that enables researchers to simulate specific reductions in gene activity. This innovative tool aids in pinpointing and comprehending crucial elements of the genome and holds the potential to empower scientists lacking access to physical labs to make significant discoveries.
Picture yourself sifting through vast numbers of unexplained genetic mutations. With CRISPR gene-editing technology, only a few of these mutations might hold therapeutic promise. However, validating these possibilities would require countless hours in a lab. Dedicating time and funds to determine which mutations merit further investigation could be quite taxing. But what if this exploration could occur in a virtual environment using artificial intelligence?
CREME is a cutting-edge AI-based virtual lab created by Assistant Professor Peter Koo and his team from Cold Spring Harbor Laboratory (CSHL). This tool allows geneticists to conduct thousands of virtual trials instantly. Scientists can now utilize it to start identifying and understanding vital areas of the genome.
The platform is designed around CRISPR interference (CRISPRi), a technique that modifies gene expression using CRISPR. CRISPRi enables biologists to decrease the activity of certain genes within a cell. In a similar vein, CREME allows researchers to make virtual adjustments to the genome and forecast how these will affect gene function, effectively providing an AI-driven version of CRISPRi.
“In practice, performing CRISPRi in a lab is immensely complex. You’re also constrained by the number of alterations and the scale of your experiments. However, since our alterations are conducted virtually, we can expand our experimental range significantly. The volume of experiments we’ve been able to conduct is unmatched – hundreds of thousands of perturbation tests,” Koo explains.
Koo and his team evaluated CREME alongside another AI-based genome analysis tool known as Enformer. Their goal was to understand how the algorithm of Enformer makes genomic predictions, which is a crucial aspect of Koo’s research focus.
“We possess these substantial and powerful models capable of accurately predicting gene expression from DNA sequences. However, understanding how these models derive their predictions has been a challenge. It’s likely they’ve identified many of the underlying principles of gene regulation, but we are still uncertain about the basis of their predictions,” Koo states.
Using CREME, Koo’s team identified a variety of genetic principles that Enformer recognized while examining the genome. This knowledge could be invaluable for the future of drug discovery. “Grasping the principles of gene regulation provides you with greater flexibility in adjusting gene expression levels in specific and predictable manners,” Koo remarks.
With additional refinement, CREME could potentially guide geneticists towards the discovery of new therapeutic targets. Most importantly, it could empower scientists without access to traditional laboratories to achieve these pioneering insights.
