iDEC 2024 | STU -China

Background:

Extremophiles, known for their ability to thrive in extreme environments, are valuable for various applications ranging from material synthesis to environmental monitoring. However, identifying these microorganisms has traditionally been a laborious process, heavily reliant on manual screening methods. To address this challenge, the need for more efficient and accurate identification techniques has become essential.

Results:

In response, we developed iExtreme, a comprehensive database containing 1,030 genomes of extremophiles across three categories, alongside a deep learning method for their identification. This method achieved an accuracy of up to 0.99 in predicting extremophile living conditions. Through iExtreme, we identified 520 previously unknown extremophilic species and 4,419 extremophilic genomes from various databases. Additionally, we utilized structure-based clustering to discover novel D-psicose 3-epimerases (DPEase) and α-amylases. To further enhance enzyme activity, we developed a directed evolution method using phage-assisted non-continuous evolution in droplets. The evolved DPEase demonstrated a 3-fold increase in activity, a 5.4-fold extension in half-life, and achieved the highest reported yield of 243 g/L D-allulose.

---