Michael Hu

Workshop title:
Prompt Bioinformatics with ChatGPT


In this workshop I will introduce the concept of prompt bioinformatics, namely the use of natural language instruction to guide chatbots for bioinformatics coding. ChatGPT is knowledgeable in both biology and informatics, exhibiting the potential of being a virtual assistant for learning bioinformatics. However, the accuracy of response depends heavily on the quality of prompts. To this end, we introduce the OPTIMAL model to guide communication with chatbots through iterative mentoring and assessment for prompt-optimization. Then we will explore a couple of online tools that make the OPTIMAL practice possible. Lastly, we will briefly touch base on ChatGPT’s new feature on image reading and assessing its capability in interpreting bioinformatics figures.