양적분자생물학 연구실

We want to know how genes are regulated. Most of all, we are drawn to what happens after transcription, when RNA-binding proteins settle what becomes of each mRNA. It is hard to watch, so we build the methods to see it: high-throughput biochemistry and sequencing that take in the whole transcriptome at once. Looking at every gene lets the data show us rules we did not know to look for. The rules for reading an mRNA are also the rules for writing one, and we use them to design mRNA vaccines and therapeutics. A single protein can be written as more than 10600 different mRNA sequences, far more than there are atoms in the universe, because most amino acids have several synonymous codons. Only some of those sequences make a good medicine. And a vaccine mRNA gets one chance. The cell keeps transcribing its own mRNAs, but the one we deliver cannot be replaced once it is gone, so how long it survives and how efficiently it is read decide how much protein it makes. Finding the right sequence starts at the bench. We measure what really happens to mRNAs in cells, turn those measurements into the rules our algorithms follow, and send the sequences they design back to the bench to be tested. Testing the rules. Design tools optimize features that are believed to matter. Which ones really do is still an open question. We answer it in cells, measuring thousands of sequence variants side by side to see what changes how long an mRNA lasts and how much protein it makes. Design algorithms. The answers go into software that has to search that astronomical space well. It weighs every shape an RNA can fold into and protects the regulatory elements in the untranslated regions at either end. It is also fast: running on parallel cloud computers, it can keep pace with an outbreak response such a