ABSTRACT
Evolution-based deep generative models represent an exciting direction in understanding and designing proteins. An open question is whether such models can learn specialized functional constraints that control fitness in specific biological contexts. Here, we examine the ability of generative models to produce synthetic versions of Src-homology 3 (SH3) domains that mediate signaling in the Sho1 osmotic stress response pathway of yeast. We show that a variational autoencoder (VAE) model produces artificial sequences that experimentally recapitulate the function of natural SH3 domains. More generally, the model organizes all fungal SH3 domains such that locality in the model latent space (but not simply locality in sequence space) enriches the design of synthetic orthologs and exposes non-obvious amino acid constraints distributed near and far from the SH3 ligand-binding site. The ability of generative models to design ortholog-like functions in vivo opens new avenues for engineering protein function in specific cellular contexts and environments.
Subject(s)
Deep Learning , Signal Transduction , src Homology Domains , Saccharomyces cerevisiae/genetics , Saccharomyces cerevisiae/metabolism , Saccharomyces cerevisiae Proteins/genetics , Saccharomyces cerevisiae Proteins/metabolismABSTRACT
Proteins are molecular machines and to understand how they work, we need to understand how they move. New pump-probe time-resolved X-ray diffraction methods open up ways to initiate and observe protein motions with atomistic detail in crystals on biologically relevant timescales. However, practical limitations of these experiments demands parallel development of effective molecular dynamics approaches to accelerate progress and extract meaning. Here, we establish robust and accurate methods for simulating dynamics in protein crystals, a nontrivial process requiring careful attention to equilibration, environmental composition, and choice of force fields. With more than seven milliseconds of sampling of a single chain, we identify critical factors controlling agreement between simulation and experiments and show that simulated motions recapitulate ligand-induced conformational changes. This work enables a virtuous cycle between simulation and experiments for visualizing and understanding the basic functional motions of proteins.
Subject(s)
Molecular Dynamics Simulation , Proteins , Proteins/metabolism , X-Ray Diffraction , Protein ConformationABSTRACT
A major goal in biomedical science is to move beyond static images of proteins and other biological macromolecules to the internal dynamics underlying their function. This level of study is necessary to understand how these molecules work and to engineer new functions and modulators of function. Stemming from a visionary commitment to this problem by Keith Moffat decades ago, a community of structural biologists has now enabled a set of x-ray scattering technologies for observing intramolecular dynamics in biological macromolecules at atomic resolution and over the broad range of timescales over which motions are functionally relevant. Many of these techniques are provided by BioCARS, a cutting-edge synchrotron radiation facility built under Moffat leadership and located at the Advanced Photon Source at Argonne National Laboratory. BioCARS enables experimental studies of molecular dynamics with time resolutions spanning from 100 ps to seconds and provides both time-resolved x-ray crystallography and small- and wide-angle x-ray scattering. Structural changes can be initiated by several methods-UV/Vis pumping with tunable picosecond and nanosecond laser pulses, substrate diffusion, and global perturbations, such as electric field and temperature jumps. Studies of dynamics typically involve subtle perturbations to molecular structures, requiring specialized computational techniques for data processing and interpretation. In this review, we present the challenges in experimental macromolecular dynamics and describe the current state of experimental capabilities at this facility. As Moffat imagined years ago, BioCARS is now positioned to catalyze the scientific community to make fundamental advances in understanding proteins and other complex biological macromolecules.
ABSTRACT
Autologous dendritic cell (DC)-based immunotherapy is a cell-based advanced therapy medicinal product (ATMP) that was first introduced more than three decades ago. In the current study, our objective was to establish a harmonized protocol using two varied antigenic sources and a good manufacturing practice (GMP)-compliant, manual method for generating clinical-grade DCs at a limited-resource academic setting. After obtaining ethical committee-approved informed consent, the recruited patients underwent leukapheresis, and single-batch DC production was carried out. Using responder-independent flow cytometric assays as quality control (QC) criteria, we propose a differentiation and maturation index (DI and MI, respectively), calculated with the QC cut-off and actual scores of each batch for comparison. Changes during cryopreservation and personnel variation were assessed periodically for up to two to three years. Using our harmonized batch production protocol, the average DI was 1.39 and MI was 1.25. Allogenic responder proliferation was observed in all patients, while IFN-gamma secretion, evaluated using flow cytometry, was detected in 10/36 patients and significantly correlated with CD8+ T cell proliferation (p value-0.0002). Tracking the viability and phenotype of cryopreserved MDCs showed a >90% viability for up to three years, while a mature DC phenotype was retained for up to one year. Our results confirm that the manual/semi-automated protocol was simple, consistent, and cost-effective, without the requirement for expensive equipment and without compromising on the quality of the final product.
ABSTRACT
Proteins are molecular machines and to understand how they work, we need to understand how they move. New pump-probe time-resolved X-ray diffraction methods open up ways to initiate and observe protein motions with atomistic detail in crystals on biologically relevant timescales. However, practical limitations of these experiments demands parallel development of effective molecular dynamics approaches to accelerate progress and extract meaning. Here, we establish robust and accurate methods for simulating dynamics in protein crystals, a nontrivial process requiring careful attention to equilibration, environmental composition, and choice of force fields. With more than seven milliseconds of sampling of a single chain, we identify critical factors controlling agreement between simulation and experiments and show that simulated motions recapitulate ligand-induced conformational changes. This work enables a virtuous cycle between simulation and experiments for visualizing and understanding the basic functional motions of proteins.
ABSTRACT
Deep generative models (DGMs) have shown great success in the understanding and data-driven design of proteins. Variational autoencoders (VAEs) are a popular DGM approach that can learn the correlated patterns of amino acid mutations within a multiple sequence alignment (MSA) of protein sequences and distill this information into a low-dimensional latent space to expose phylogenetic and functional relationships and guide generative protein design. Autoregressive (AR) models are another popular DGM approach that typically lacks a low-dimensional latent embedding but does not require training sequences to be aligned into an MSA and enable the design of variable length proteins. In this work, we propose ProtWave-VAE as a novel and lightweight DGM, employing an information maximizing VAE with a dilated convolution encoder and an autoregressive WaveNet decoder. This architecture blends the strengths of the VAE and AR paradigms in enabling training over unaligned sequence data and the conditional generative design of variable length sequences from an interpretable, low-dimensional learned latent space. We evaluated the model's ability to infer patterns and design rules within alignment-free homologous protein family sequences and to design novel synthetic proteins in four diverse protein families. We show that our model can infer meaningful functional and phylogenetic embeddings within latent spaces and make highly accurate predictions within semisupervised downstream fitness prediction tasks. In an application to the C-terminal SH3 domain in the Sho1 transmembrane osmosensing receptor in baker's yeast, we subject ProtWave-VAE-designed sequences to experimental gene synthesis and select-seq assays for the osmosensing function to show that the model enables synthetic protein design, conditional C-terminus diversification, and engineering of the osmosensing function into SH3 paralogues.
Subject(s)
Genetic Techniques , Proteins , Phylogeny , Mutation , Amino Acid SequenceABSTRACT
BACKGROUND: Population-based cancer survival is a key measurement of cancer control performance linked to diagnosis and treatment, but benchmarking studies that include lower-income settings and that link results to health systems and human development are scarce. SURVCAN-3 is an international collaboration of population-based cancer registries that aims to benchmark timely and comparable cancer survival estimates in Africa, central and south America, and Asia. METHODS: In SURVCAN-3, population-based cancer registries from Africa, central and south America, and Asia were invited to contribute data. Quality control and data checks were carried out in collaboration with population-based cancer registries and, where applicable, active follow-up was performed at the registry. Patient-level data (sex, age at diagnosis, date of diagnosis, morphology and topography, stage, vital status, and date of death or last contact) were included, comprising patients diagnosed between Jan 1, 2008, and Dec 31, 2012, and followed up for at least 2 years (until Dec 31, 2014). Age-standardised net survival (survival where cancer was the only possible cause of death), with 95% CIs, at 1 year, 3 years, and 5 years after diagnosis were calculated using Pohar-Perme estimators for 15 major cancers. 1-year, 3-year, and 5-year net survival estimates were stratified by countries within continents (Africa, central and south America, and Asia), and countries according to the four-tier Human Development Index (HDI; low, medium, high, and very high). FINDINGS: 1 400 435 cancer cases from 68 population-based cancer registries in 32 countries were included. Net survival varied substantially between countries and world regions, with estimates steadily rising with increasing levels of the HDI. Across the included cancer types, countries within the lowest HDI category (eg, CÔte d'Ivoire) had a maximum 3-year net survival of 54·6% (95% CI 33·3-71·6; prostate cancer), whereas those within the highest HDI categories (eg, Israel) had a maximum survival of 96·8% (96·1-97·3; prostate cancer). Three distinct groups with varying outcomes by country and HDI dependant on cancer type were identified: cancers with low median 3-year net survival (<30%) and small differences by HDI category (eg, lung and stomach), cancers with intermediate median 3-year net survival (30-79%) and moderate difference by HDI (eg, cervix and colorectum), and cancers with high median 3-year net survival (≥80%) and large difference by HDI (eg, breast and prostate). INTERPRETATION: Disparities in cancer survival across countries were linked to a country's developmental position, and the availability and efficiency of health services. These data can inform policy makers on priorities in cancer control to reduce apparent inequality in cancer outcome. FUNDING: Tata Memorial Hospital, the Martin-Luther-University Halle-Wittenberg, and the International Agency for Research on Cancer.
Subject(s)
Benchmarking , Prostatic Neoplasms , Male , Female , Humans , Breast , Income , Africa, Central , RegistriesABSTRACT
Protein structure, function, and evolution depend on local and collective epistatic interactions between amino acids. A powerful approach to defining these interactions is to construct models of couplings between amino acids that reproduce the empirical statistics (frequencies and correlations) observed in sequences comprising a protein family. The top couplings are then interpreted. Here, we show that as currently implemented, this inference unequally represents epistatic interactions, a problem that fundamentally arises from limited sampling of sequences in the context of distinct scales at which epistasis occurs in proteins. We show that these issues explain the ability of current approaches to predict tertiary contacts between amino acids and the inability to obviously expose larger networks of functionally relevant, collectively evolving residues called sectors. This work provides a necessary foundation for more deeply understanding and improving evolution-based models of proteins.
Subject(s)
Amino Acids , Proteins , Proteins/metabolismABSTRACT
Metaplastic carcinoma (MPC) is a rare subgroup of breast tumours accounting for <5% of all invasive breast cancers. Histologically confirmed 40 MPC from January 2001 to December 2018 were identified from our electronic database: stage I 2.5% (n = 1), stage II 40% (n = 16), stage III 45% (n = 18) and stage IV 12.5% (n = 5). The mean tumour size was 6 cm, node-negative in 60%, and hormone receptor-negative in 75%. Among the 35 non-metastatic patients, 17 (48.6%) received initial neoadjuvant treatment (NAT), followed by surgery, and only 1 had a complete pathological response. At a median follow-up of 60 months, 17% (n = 6) had a recurrence. All six of them had lung metastasis. The 5-year overall survival (OS) and disease-free survival were 64.4% and 66.3%, respectively. Age more than 46 years (p = 0.027), tumour size more than 5 cm (p = 0.037), and nodal positivity (p = 0.001) were predictors of OS. In node-positive patients, the 5-year OS in those who underwent initial surgery was 80% and after NAT was 21.4% (p = 0.069). In node-negative patients, the 5-year OS after initial surgery was 83.3% and after NAT was 90% (p = 0.380). A statistical significance could not be demonstrated due to the small number of patients. Due to chemoresistance, the concept of initial NAT in MPC of the breast is a subject to be studied in the future. Upfront surgery should be considered for operable diseases (including stage III), followed by a decision on adjuvant therapy. Optimal treatment and effective systemic therapy regimens are yet to be defined.
ABSTRACT
BACKGROUND AND AIM: Definitive chemoradiation (dCRT) is the standard treatment for locally advanced inoperable esophageal cancer. The aim of this study is to analyze the effect of dCRT combined with paclitaxel and carboplatin (TC) against cisplatin (CDDP) with radiation. METHODS: The study population included patients with locally advanced inoperable esophageal cancer seeking treatment at our center from March 2013 to December 2017. Case records from 66 patients were extracted. The toxicity profile of patients who received TC or CDDP was reported and analyzed. A Chi-square test and students t-test were used to analyze the categorical, and the continuous variables, respectively. The KaplanMeier method was used to estimate the survival probability. A log-rank test was applied to compare the survival differences between the two groups. RESULTS: The overall survival (OS) did not differ at 3 years between the TC and CDDP (p = 0.286). The median survival duration was 13 months for CDDP and 18 months for TC. The toxicity profile like emesis (93% CDDP vs. 25% TC), neutropenia (79% CDDP vs. 13% TC), thrombocytopenia (10% CDDP vs. 17% TC) and dyselectrolytemia (71% CDDP vs. 8% TC) were compared between the two treatment groups and found to be more in CDDP group. CONCLUSION: The treatment of patients with locally advanced esophageal carcinoma with dCRT and TC showed an improved toxicity profile, but similar OS compared to CDDP. Applying dCRT with TC could be an alternate regimen for locally advanced inoperable esophageal cancer patients. RELEVANCE FOR PATIENTS: Concurrent chemoradiation with TC regimen can be considered as an alternative for cisplatin as it shows equivalent survival and reduced toxicity profile.
ABSTRACT
Biological organisms experience constantly changing environments, from sudden changes in physiology brought about by feeding, to the regular rising and setting of the Sun, to ecological changes over evolutionary timescales. Living organisms have evolved to thrive in this changing world but the general principles by which organisms shape and are shaped by time varying environments remain elusive. Our understanding is particularly poor in the intermediate regime with no separation of timescales, where the environment changes on the same timescale as the physiological or evolutionary response. Experiments to systematically characterize the response to dynamic environments are challenging since such environments are inherently high dimensional. This roadmap deals with the unique role played by time varying environments in biological phenomena across scales, from physiology to evolution, seeking to emphasize the commonalities and the challenges faced in this emerging area of research.
Subject(s)
Biological Evolution , Environment , Physiological Phenomena , Time FactorsABSTRACT
The design of synthetic proteins with the desired function is a long-standing goal in biomolecular science, with broad applications in biochemical engineering, agriculture, medicine, and public health. Rational de novo design and experimental directed evolution have achieved remarkable successes but are challenged by the requirement to find functional "needles" in the vast "haystack" of protein sequence space. Data-driven models for fitness landscapes provide a predictive map between protein sequence and function and can prospectively identify functional candidates for experimental testing to greatly improve the efficiency of this search. This Viewpoint reviews the applications of machine learning and, in particular, deep learning as part of data-driven protein engineering platforms. We highlight recent successes, review promising computational methodologies, and provide an outlook on future challenges and opportunities. The article is written for a broad audience comprising both polymer and protein scientists and computer and data scientists interested in an up-to-date review of recent innovations and opportunities in this rapidly evolving field.
Subject(s)
Anniversaries and Special Events , Proteins , Amino Acid Sequence , Machine Learning , Macromolecular Substances , Protein Engineering , Proteins/chemistryABSTRACT
The ribosome translates the genetic code into proteins in all domains of life. Its size and complexity demand long-range interactions that regulate ribosome function. These interactions are largely unknown. Here, we apply a global coevolution method, statistical coupling analysis (SCA), to identify coevolving residue networks (sectors) within the 23S ribosomal RNA (rRNA) of the large ribosomal subunit. As in proteins, SCA reveals a hierarchical organization of evolutionary constraints with near-independent groups of nucleotides forming physically contiguous networks within the three-dimensional structure. Using a quantitative, continuous-culture-with-deep-sequencing assay, we confirm that the top two SCA-predicted sectors contribute to ribosome function. These sectors map to distinct ribosome activities, and their origins trace to phylogenetic divergences across all domains of life. These findings provide a foundation to map ribosome allostery, explore ribosome biogenesis, and engineer ribosomes for new functions. Despite differences in chemical structure, protein and RNA enzymes appear to share a common internal logic of interaction and assembly.
Subject(s)
Escherichia coli/genetics , RNA, Bacterial/chemistry , RNA, Ribosomal, 23S/chemistry , Ribosomes/genetics , Escherichia coli/chemistry , Escherichia coli/metabolism , Evolution, Molecular , Nucleic Acid Conformation , Phylogeny , RNA, Bacterial/genetics , RNA, Bacterial/metabolism , RNA, Ribosomal, 23S/genetics , RNA, Ribosomal, 23S/metabolism , Ribosomes/chemistry , Ribosomes/metabolismABSTRACT
The rational design of enzymes is an important goal for both fundamental and practical reasons. Here, we describe a process to learn the constraints for specifying proteins purely from evolutionary sequence data, design and build libraries of synthetic genes, and test them for activity in vivo using a quantitative complementation assay. For chorismate mutase, a key enzyme in the biosynthesis of aromatic amino acids, we demonstrate the design of natural-like catalytic function with substantial sequence diversity. Further optimization focuses the generative model toward function in a specific genomic context. The data show that sequence-based statistical models suffice to specify proteins and provide access to an enormous space of functional sequences. This result provides a foundation for a general process for evolution-based design of artificial proteins.
Subject(s)
Chorismate Mutase , Evolution, Molecular , Models, Genetic , Models, Statistical , Amino Acid Sequence , Chorismate Mutase/chemistry , Chorismate Mutase/genetics , Escherichia coli Proteins/chemistry , Escherichia coli Proteins/geneticsABSTRACT
BACKGROUND: Breakthrough chemotherapy-induced vomiting (CIV) is defined as CIV occurring after adequate antiemetic prophylaxis. Olanzapine and metoclopramide are two drugs recommended for the treatment of breakthrough CIV in children, without adequate evidence. We conducted an open-label, single-center, phase 3 randomized controlled trial comparing the safety and efficacy of olanzapine and metoclopramide for treating breakthrough CIV. PROCEDURE: Children aged 5-18 years who developed breakthrough CIV after receiving highly emetogenic chemotherapy or moderately emetogenic chemotherapy were randomly assigned to the metoclopramide or olanzapine arm. The primary objective of the study was to compare the complete response (CR) rates between patients receiving olanzapine or metoclopramide for treating breakthrough CIV during 72 hours after the administration of the study drug. Secondary objectives were to compare CR rates for nausea and toxicities between the two arms. RESULTS: Eighty patients were analyzed (39 in the olanzapine arm and 41 in the metoclopramide arm). CR rates were significantly higher in the olanzapine arm compared with the metoclopramide arm for vomiting (72% vs 39%, P = 0.003) and nausea (59% vs 34%, P = 0.026). Seven patients in the metoclopramide arm crossed over to the olanzapine arm and none crossed over in the olanzapine arm (P < 0.001). The mean nausea score in the olanzapine arm was significantly lower than the metoclopramide arm after the initiation of the rescue antiemetic (P = 0.01). Hyperglycemia and drowsiness were more commonly seen in the olanzapine arm. CONCLUSION: Olanzapine is superior to metoclopramide for the treatment of breakthrough CIV in children. Drowsiness and hyperglycemia need to be monitored closely in children receiving olanzapine for breakthrough CIV.
Subject(s)
Antiemetics/therapeutic use , Antineoplastic Combined Chemotherapy Protocols/adverse effects , Metoclopramide/therapeutic use , Neoplasms/drug therapy , Olanzapine/therapeutic use , Vomiting/drug therapy , Adolescent , Child , Child, Preschool , Female , Follow-Up Studies , Humans , Male , Neoplasms/pathology , Prognosis , Vomiting/chemically induced , Vomiting/pathologyABSTRACT
INTRODUCTION: The aim of the study was to analyze the various prognostic factors that influence survival and clinical outcomes in patients undergoing liver resection for huge hepatocellular carcinomas. MATERIALS AND METHODS: The records of patients who underwent curative surgery between 1991 and 2011 for huge hepatocellular carcinoma were analyzed. Various prognostic factors that influenced the survival were studied. The patients were followed up till November 2016. RESULTS: The number of patients who underwent liver resection with huge hepatocellular carcinoma during the study period was 17; this included 14 males and 3 females. The median age of the study population was 52 years. The median serum AFP in the study population was 132.3 ng/ml (range 2 to 187,000 ng/ml). 41.2% of the patients were hepatitis B positive. The overall morbidity was 6%. The mortality rate was nil. The mean size of the resected specimen was 13.9 cm ± 3.6 cm. The overall recurrence rate was 76.5%. The local recurrence rate was 29.4%. The median time to recurrence was 8 months. The 5-year disease-free survival and overall survival of the study group were 26% and 32%, respectively. The factors that predicted an adverse survival outcome after the log-rank test for univariate analysis using life-table method were presence of lymphovascular invasion (p = 0.047), age ≤ 55 years (p = 0.021), and raised serum AFP (p = 0.041). CONCLUSION: The factors that predict an adverse outcome after surgery in patients with huge hepatocellular carcinomas were the presence of lymphovascular invasion, raised serum AFP, and age ≤ 55 years.
Subject(s)
Carcinoma, Hepatocellular/surgery , Liver Neoplasms/surgery , Carcinoma, Hepatocellular/mortality , Carcinoma, Hepatocellular/pathology , Female , Humans , Liver Neoplasms/mortality , Liver Neoplasms/pathology , Male , Middle Aged , Prognosis , Survival Analysis , Treatment OutcomeABSTRACT
Understanding the pattern of epistasis-the non-independence of mutations-is critical for relating genotype and phenotype. However, the combinatorial complexity of potential epistatic interactions has severely limited the analysis of this problem. Using new mutational approaches, we report a comprehensive experimental study of all 213 mutants that link two phenotypically distinct variants of the Entacmaea quadricolor fluorescent protein-an opportunity to examine epistasis up to the 13th order. The data show the existence of many high-order epistatic interactions between mutations, but also reveal extraordinary sparsity, enabling novel experimental and computational strategies for learning the relevant epistasis. We demonstrate that such information, in turn, can be used to accurately predict phenotypes in practical situations where the number of measurements is limited. Finally, we show how the observed epistasis shapes the solution space of single-mutation trajectories between the parental fluorescent proteins, informative about the protein's evolutionary potential. This work provides conceptual and experimental strategies to profoundly characterize epistasis in a protein, relevant to both natural and laboratory evolution.
Subject(s)
Epistasis, Genetic , Proteins/genetics , Evolution, Molecular , Genotype , Mutation , Phenotype , Proteins/chemistryABSTRACT
This year, the Nobel Prize in Chemistry was awarded to three pioneering scientists who applied laboratory evolution for protein engineering: Frances Arnold, George P. Smith, and Sir Gregory P. Winter. This approach has had major impact in various applications and inspires the search for the general principles of design through evolution.
Subject(s)
Directed Molecular Evolution , Nobel Prize , Protein Engineering , Directed Molecular Evolution/methods , Directed Molecular Evolution/trends , Humans , Protein Engineering/methods , Protein Engineering/trendsABSTRACT
The functional diversity of kinases enables specificity in cellular signal transduction. Yet how more than 500 members of the human kinome specifically receive regulatory inputs and convey information to appropriate substrates-all while using the common signaling output of phosphorylation-remains enigmatic. Here, we perform statistical co-evolution analysis, mutational scanning, and quantitative live-cell assays to reveal a hierarchical organization of the kinase domain that facilitates the orthogonal evolution of regulatory inputs and substrate outputs while maintaining catalytic function. We find that three quasi-independent "sectors"-groups of evolutionarily coupled residues-represent functional units in the kinase domain that encode for catalytic activity, substrate specificity, and regulation. Sector positions impact both disease and pharmacology: the catalytic sector is significantly enriched for somatic cancer mutations, and residues in the regulatory sector interact with allosteric kinase inhibitors. We propose that this functional architecture endows the kinase domain with inherent regulatory plasticity.
Subject(s)
Catalytic Domain , Evolution, Molecular , Protein Kinases/chemistry , Allosteric Regulation , Allosteric Site , Humans , Mutation , Neoplasms/genetics , Protein Kinases/genetics , Protein Kinases/metabolism , Substrate SpecificityABSTRACT
Protein function arises from a poorly understood pattern of energetic interactions between amino acid residues. Sequence-based strategies for deducing this pattern have been proposed, but lack of benchmark data has limited experimental verification. Here, we extend deep-mutation technologies to enable measurement of many thousands of pairwise amino acid couplings in several homologs of a protein family - a deep coupling scan (DCS). The data show that cooperative interactions between residues are loaded in a sparse, evolutionarily conserved, spatially contiguous network of amino acids. The pattern of amino acid coupling is quantitatively captured in the coevolution of amino acid positions, especially as indicated by the statistical coupling analysis (SCA), providing experimental confirmation of the key tenets of this method. This work exposes the collective nature of physical constraints on protein function and clarifies its link with sequence analysis, enabling a general practical approach for understanding the structural basis for protein function.