3 papers
cs.LG2025
In-Context Learning can distort the relationship between sequence likelihoods and biological fitness
Pranav Kantroo, Günter P. Wagner, Benjamin B. Machta
Language models have emerged as powerful predictors of the viability of biological sequences. During training these models learn the rules of the grammar obeyed by sequences of ami…
q-bio.BM2024
High fitness paths can connect proteins with low sequence overlap
Pranav Kantroo, Günter P. Wagner, Benjamin B. Machta
The structure and function of a protein are determined by its amino acid sequence. While random mutations change a protein's sequence, evolutionary forces shape its structural fold…
q-bio.BM2024
Pseudo-perplexity in One Fell Swoop for Protein Fitness Estimation
Pranav Kantroo, Günter P. Wagner, Benjamin B. Machta
Protein language models trained on the masked language modeling objective learn to predict the identity of hidden amino acid residues within a sequence using the remaining observab…