Biology subjects

Martinez, Z. A.

Publications and source records attributed to Martinez, Z. A..

5 recordsLinked to original sources

bioRxiv2025

Pretrained protein language models choose between sequence novelty and structural completeness

Protein language models (PLMs) have gained increasing acceptance in tasks ranging from variant effect prediction in disease to optimization and de novo design of proteins with improved stability, target-binding affinity, and catalytic performance. Despite encouraging performance in such applications, little is understood as far as the degree to which PLM-generated sequences - putative novel protein outputs - recapitulate the broad biophysical rules and diversity of sequence, structure, and function that defines natural protein-space, vital knowledge for boosting the design capacity of PLMs in ever-more-complex systems. Towards this end, we computationally profile and characterize the sequence and structure statistics and properties of hundreds of thousands of potential small proteins proposed through free unconstrained generation from architecturally distinct PLMs. We show that although these models exhibit a prodigious latent capacity to access novel amino-acid sequences, they struggle to approach the structural variation that exists on plain display in nature. Moreover, we uncover a stark tradeoff between prioritizing sequence novelty or structural breadth, exemplified by a "helical bundle trap" that dominates model output when aiming outside the comfortable bounds and evolutionary organization of natural sequences. These findings underscore a critical need for strategies that can rapidly guide PLMs into unlocking through generation the full richness of protein sequence, structure, and function that is consistent with governing biophysics but tantalizingly untapped as of yet in design contexts. Author summaryLarge language models (LLMs) like GPT arent just for human text. Spinoff versions that treat protein and DNA sequences as special "languages" of their own, complete with preferred words and grammars, are being used to identify disease-causing genes and mutations and to design new treatments and drugs for clinical testing. But as anyone who has used an LLM chatbot has probably experienced at one time or another, these models can act narrow-minded or nonsensical, reducing their generality and utility. Protein language models are no exception to such flaws of reasoning. We show that protein language models face a fundamental choice between suggesting novel sequences that look nothing like natural ones and capturing the full range of three-dimensional shapes and structures responsible for the diverse functioning of molecular machines. Managing or bypassing this tradeoff is consequently of high import for designing proteins that impart novel functions and activities for therapeutic targeting and beyond.

bioinformatics↗

bioRxiv2024

Notch signaling stabilizes lengths of motile cilia in multiciliated cells in the lung

Airway multiciliated cells (MCs) maintain respiratory health by clearing mucus and trapped particles through the beating of motile cilia. While it is known that ciliary lengths decrease along the proximal-distal (P-D) axis of the tracheobronchial tree, how this is regulated is unclear. Here, we demonstrate that canonical Notch signaling in MCs plays a critical role in stabilizing ciliary length. Inhibition of Notch signaling in MCs results in ciliary shortening in the trachea, lengthening in the distal airway, and to region-specific alterations in gene expression. We probe how environmental challenges impact MC homeostasis using germ-free and Mycobacterium tuberculosis (M. tb) infection models. While germ-free conditions do not perturb ciliary lengths, M. tb infection leads to lengthening of distal airway cilia, correlating with a downregulation of Notch signaling. These findings reveal that ciliary length and the P-D gradient in the airways are actively regulated, with Notch signaling serving as a stabilizing mechanism.

cell biology↗

bioRxiv2024

Benchmarking Protein Language Models for Protein Crystallization

The problem of protein structure determination is usually solved by X-ray crystallography. Several in silico deep learning methods have been developed to overcome the high attrition rate, cost of experiments and extensive trial-and-error settings, for the predicting the crystallization propensities of proteins based on their sequences. In this work, we benchmark the power of open protein language models (PLMs) through the TRILL platform, a bespoke framework democratizing the usage of PLMs for the task of predicting crystallization propensities of proteins. By comparing LightGBM / XGBoost classifiers built on the embedding representations learned by different PLMs, such as ESM2, Ankh, ProtT5-XL, ProstT5, with the performance of state-of-the-art sequence-based methods like DeepCrystal, ATTCrys and CLPred, we identify the most effective methods for predicting crystallization outcomes. The LightGBM classifiers utilizing embeddings from ESM2 model with 30 and 36 transformer layers and 150 and 3, 000 million parameters respectively have performance gains by 3 - 5% then all compared models for various evaluation metrics, including AUPR (Area Under Precision-Recall Curve), AUC (Area Under the Receiver Operating Characteristic Curve), and F1 score on independent test sets. Furthermore, we fine-tune the ProtGPT2 model available via TRILL to generate crystallizable proteins. Starting with 3, 000 generated proteins and through a step of filtration processes including consensus of all open PLM-based classifiers, sequence identity through CD-HIT, secondary structure compatibility, aggregation screening, homology search and foldability evaluation, we identified a set of 5 novel proteins as potentially crystallizable.

bioinformatics↗

bioRxiv2024

Optimizing protein expression in the One-Pot PURE system: insights into reaction composition and translation efficiency

The One-Pot PURE (Protein synthesis Using Recombinant Elements) system simplifies the preparation of traditional PURE systems by co-culturing and purifying 36 essential proteins for gene expression in a single step, thereby improving accessibility and affordability for widespread laboratory adoption and customization. However, replicating this protocol to match the productivity of traditional PURE systems can take considerable time and effort due to uncharacterized variability in the systems biochemical composition. In this work, we observed unstable PURE protein expression in E. coli strains M15/pREP4 and BL21(DE3) and addressed this using glucose-mediated catabolite repression to minimize burdensome background expression. We also identified differences in optimal protein induction timing between these two strains, leading to growth incompatibility in co-culture, and observed proteolysis of PURE proteins expressed in M15/pREP4. We showed that consolidating all expression vectors into a protease-deficient BL21(DE3) strain could minimize proteolysis. This single-strain system also led to more uniform cell growth at the time of protein induction, improving the stoichiometry of critical translation initiation factors in the PURE reaction for efficient protein production. In addition to optimizing One-Pot PURE protein composition, we found that variations in commercial energy solution formulations could compensate for suboptimal PURE protein stoichiometry. Moreover, our study revealed significant differences in the expression capacity of commercially available E. coli tRNAs, suggesting the potential of optimizing tRNA codons to improve protein translation. Taken together, this work highlights the complex biochemical interplay influencing protein expression capacity in the One-Pot PURE system and presents strategies to improve its robustness and productivity.

synthetic biology↗

bioRxiv2023

TRILL: Orchestrating Modular Deep-Learning Workflows for Democratized, Scalable Protein Analysis and Engineering

AO_SCPLOWBSTRACTC_SCPLOWDeep-learning models have been rapidly adopted by many fields, partly due to the deluge of data humanity has amassed. In particular, the petabases of biological sequencing data enable the unsupervised training of protein language models that learn the "language of life." However, due to their prohibitive size and complexity, contemporary deep-learning models are often unwieldy, especially for scientists with limited machine learning backgrounds. TRILL (TRaining and Inference using the Language of Life) is a platform for creative protein design and discovery. Leveraging several state-of-the-art models such as ESM-2, DiffDock, and RFDiffusion, TRILL allows researchers to generate novel proteins, predict 3-D structures, extract high-dimensional representations of proteins, functionally classify proteins and more. What sets TRILL apart is its ability to enable complex pipelines by chaining together models and effectively merging the capabilities of different models to achieve a sum greater than its individual parts. Whether using Google Colab with one GPU or a supercomputer with hundreds, TRILL allows scientists to effectively utilize models with millions to billions of parameters by using optimized training strategies such as ZeRO-Offload and distributed data parallel. Therefore, TRILL not only bridges the gap between complex deep-learning models and their practical application in the field of biology, but also simplifies the orchestration of these models into comprehensive workflows, democratizing access to powerful methods. Documentation: https://trill.readthedocs.io/en/latest/home.html.

bioengineering↗