Sequence and structural conservation reveal fingerprint residues in TRP channels.

Transient receptor potential (TRP) proteins are a large family of cation-selective channels, surpassed in variety only by voltage-gated potassium channels. Detailed molecular mechanisms governing how membrane voltage, ligand binding, or temperature can induce conformational changes promoting the open state in TRP channels are still a matter of debate. Aiming to unveil distinctive structural features common to the transmembrane domains within the TRP family, we performed phylogenetic reconstruction, sequence statistics, and structural analysis over a large set of TRP channel genes. Here, we report an exceptionally conserved set of residues. This fingerprint is composed of twelve residues localized at equivalent three-dimensional positions in TRP channels from the different subtypes. Moreover, these amino acids are arranged in three groups, connected by a set of aromatics located at the core of the transmembrane structure. We hypothesize that differences in the connectivity between these different groups of residues harbor the apparent differences in coupling strategies used by TRP subgroups.

The generative capacity of probabilistic protein sequence models.

Potts models and variational autoencoders (VAEs) have recently gained popularity as generative protein sequence models (GPSMs) to explore fitness landscapes and predict mutation effects. Despite encouraging results, current model evaluation metrics leave unclear whether GPSMs faithfully reproduce the complex multi-residue mutational patterns observed in natural sequences due to epistasis. Here, we develop a set of sequence statistics to assess the "generative capacity" of three current GPSMs: the pairwise Potts Hamiltonian, the VAE, and the site-independent model. We show that the Potts model's generative capacity is largest, as the higher-order mutational statistics generated by the model agree with those observed for natural sequences, while the VAE's lies between the Potts and site-independent models. Importantly, our work provides a new framework for evaluating and interpreting GPSM accuracy which emphasizes the role of higher-order covariation and epistasis, with broader implications for probabilistic sequence models in general.