Davide De Lucrezia

Coping with complexity: Machine learning optimization of cell‐free protein synthesis

Biological systems contain complex metabolic pathways with many nonlinearities and synergies that make them difficult to predict from first principles. Protein synthesis is a canonical example of such a pathway. Here we show how cell‐free protein synthesis may be improved through a series of iterated high‐throughput experiments guided by a machine‐learning algorithm implementing a form of evolutionary design of experiments (Evo‐DoE). The algorithm predicts fruitful experiments from statistical models of the previous experimental results, combined with stochastic exploration of the experimental space. The desired experimental response, or evolutionary fitness, was defined as the yield of the target product, and new experimental conditions were discovered to have ∼350% greater yield than the standard. An analysis of the best experimental conditions discovered indicates that there are two distinct classes of kinetics, thus showing how our evolutionary design of experiments is capable of significant innovation, as well as gradual improvement. Biotechnol. Bioeng. 2011;108:2218–2228.

Defined DNA-mediated assemblies of gene-expressing giant unilamellar vesicles

The technological aspects of artificial vesicles as prominent cell mimics are evolving toward higher-order assemblies of functional vesicles with tissuelike architectures. Here, we demonstrate the spatially controlled DNA-directed bottom-up synthesis of complex microassemblies and macroassemblies of giant unilamellar vesicles functionalized with a basic cellular machinery to express green fluorescent protein and specified neighbor-to-neighbor interactions. We show both that the local and programmable DNA pairing rules on the nanoscale are able to direct the microscale vesicles into macroscale soft matter assemblies and that the highly sensitive gene-expression machinery remains intact and active during multiple experimental steps. An in silico model recapitulates the experiments performed in vitro and covers additional experimental setups highlighting the parameters that control the DNA-directed bottom-up synthesis of higher-order self-assembled structures. The controlled assembly of a functional vesicle matrix may be useful not only as simplified natural tissue mimics but also as artificial scaffolds that could interact and support living cells.

A stochastic model of the emergence of autocatalytic cycles

Autocatalytic cycles are rather common in biological systems and they might have played a major role in the transition from non-living to living systems. Several theoretical models have been proposed to address the experimentalists during the investigation of this issue and most of them describe a phase transition depending upon the level of heterogeneity of the chemical soup. Nevertheless, it is well known that reproducing the emergence of autocatalytic sets in wet laboratories is a hard task. Understanding the rationale at the basis of such a mismatch between theoretical predictions and experimental observations is therefore of fundamental importance.We here introduce a novel stochastic model of catalytic reaction networks, in order to investigate the emergence of autocatalytic cycles, sensibly considering the importance of noise, of small-number effects and the possible growth of the number of different elements in the system.Furthermore, the introduction of a temporal threshold that defines how long a specific reaction is kept in the reaction graph allows to univocally define cycles also within an asynchronous framework.The foremost analyses have been focused on the study of the variation of the composition of the incoming flux. It was possible to show that the activity of the system is enhanced, with particular regard to the emergence of autocatalytic sets, if a larger number of different elements is present in the incoming flux, while the specific length of the species seems to entail minor effects on the overall dynamics.

Do Natural Proteins Differ from Random Sequences Polypeptides? Natural vs. Random Proteins Classification Using an Evolutionary Neural Network

Are extant proteins the exquisite result of natural selection or are they random sequences slightly edited by evolution? This question has puzzled biochemists for long time and several groups have addressed this issue comparing natural protein sequences to completely random ones coming to contradicting conclusions. Previous works in literature focused on the analysis of primary structure in an attempt to identify possible signature of evolutionary editing. Conversely, in this work we compare a set of 762 natural proteins with an average length of 70 amino acids and an equal number of completely random ones of comparable length on the basis of their structural features. We use an ad hoc Evolutionary Neural Network Algorithm (ENNA) in order to assess whether and to what extent natural proteins are edited from random polypeptides employing 11 different structure-related variables (i.e. net charge, volume, surface area, coil, alpha helix, beta sheet, percentage of coil, percentage of alpha helix, percentage of beta sheet, percentage of secondary structure and surface hydrophobicity). The ENNA algorithm is capable to correctly distinguish natural proteins from random ones with an accuracy of 94.36%. Furthermore, we study the structural features of 32 random polypeptides misclassified as natural ones to unveil any structural similarity to natural proteins. Results show that random proteins misclassified by the ENNA algorithm exhibit a significant fold similarity to portions or subdomains of extant proteins at atomic resolution. Altogether, our results suggest that natural proteins are significantly edited from random polypeptides and evolutionary editing can be readily detected analyzing structural features. Furthermore, we also show that the ENNA, employing simple structural descriptors, can predict whether a protein chain is natural or random.

Investigation of de novo totally random biosequences Part III RNA Foster : A novel assay to investigate RNA folding structural properties

Fold is essential to RNA properties, and, in particular, its thermodynamic stability can be used to monitor RNA–protein or RNA–ligand interactions, and to engineer RNA with novel or improved properties. While clearly valuable, experimental determination of RNA folding stability by traditional biophysical techniques requires substantial amounts of pure sample and rather expensive equipment. In this paper, we report a new, simple approach to the determination of RNA folding stability by coupling enzymatic digestion and temperature denaturation. The assay, named RNA folding stability Test (RNA Foster), is designed to probe the fraction of folded RNA (ffold) in an equilibrium mixture of folded and unfolded ones as a function of temperature. The simplicity of RNA Foster suggests that it can easily be scaled up for high‐throughput studies of RNA folding stability both in basic and applied research.

Question 5: On the Chemical Reality of the RNA World

The discovery of catalytic RNA has revolutionised modern molecular biology and bears important implications for the origin of Life research. Catalytic RNA, in particular self-replicating RNA, prompted the hypothesis of an early “RNA world” where RNA molecules played all major roles such information storage and catalysis. The actual role of RNA as primary actor in the origin of life has been under debate for a long time, with a particular emphasis on possible pathways to the prebiotic synthesis of mononucleotides; their polymerization and the possibility of spontaneous emergence of catalytic RNAs synthesised under plausible prebiotic conditions. However, little emphasis has been put on the chemical reality of an RNA world; in particular concerning the chemical constrains that such scenario should have met to be feasible. This paper intends to address those concerns with regard to the achievement of high local RNA molecules concentration and the aetiology of unique sequence under plausible prebiotic conditions.

Selection Dynamic of Escherichia coli Host in M13 Combinatorial Peptide Phage Display Libraries

Phage display relies on an iterative cycle of selection and amplification of random combinatorial libraries to enrich the initial population of those peptides that satisfy a priori chosen criteria. The effectiveness of any phage display protocol depends directly on library amino acid sequence diversity and the strength of the selection procedure. In this study we monitored the dynamics of the selective pressure exerted by the host organism on a random peptide library in the absence of any additional selection pressure. The results indicate that sequence censorship exerted by Escherichia coli dramatically reduces library diversity and can significantly impair phage display effectiveness.

Question 3: The Worlds of the Prebiotic and Never Born Proteins

Starting from the statement that no reliable methods are known to produce high molecular weight polypeptides under prebiotic conditions, a possible approach, at least to understand the differences between extant proteins and the possible large number of never born proteins, could be biological. Using the phage display method a large library of totally random amino acidic sequences was obtained. Consequently, different experiments to directly consider the frequency of stable folds were performed, and the interesting results obtained from such new approach are discussed in terms of contingency, contributing to the discussion on the selection mechanism of extant proteins.

Optimization of thermophilic trans‐isoprenyl diphosphate synthase expression in Escherichia coli by response surface methodology

We optimized the heterologous expression of trans‐isoprenyl diphosphate synthase (IDS), the key enzyme involved in the biosynthesis of trans‐polyisoprene. trans‐Polyisoprene is a particularly valuable compound due to its superior stiffness, excellent insulation, and low thermal expansion coefficient. Currently, trans‐polyisoprene is mainly produced through chemical synthesis and no biotechnological processes have been established so far for its large‐scale production. In this work, we employed D‐optimal design and response surface methodology to optimize the expression of thermophilic enzymes IDS from Thermococcus kodakaraensis. The design of experiment took into account of six factors (preinduction cell density, inducer concentration, postinduction temperature, salt concentration, alternative carbon source, and protein inhibitor) and seven culture media (LB, NZCYM, TB, M9, Ec, Ac, and EDAVIS) at five different pH points. By screening only 109 experimental points, we were able to improve IDS production by 48% in close‐batch fermentation.

Nave Bayes ant colony optimization for designing high dimensional experiments

Graphical abstractDisplay Omitted HighlightsA novel design optimization technique is proposed based on the ant colony optimization (ACO) framework.The proposed approach combines the properties and peculiarities of ACO and Nave Bayes classifier.The proposed approach is applied to a problem concerning enzyme engineering.The simulation and experimental results demonstrate a remarkable scalability of the proposed approach to the overall size of the search space. In a large number of experimental problems, high dimensionality of the search area and economical constraints can severely limit the number of experimental points that can be tested. Within these constraints, classical optimization techniques perform poorly, in particular, when little a priori knowledge is available. In this work we investigate the possibility of combining approaches from statistical modeling and bio-inspired algorithms to effectively explore a huge search space, sampling only a limited number of experimental points. To this purpose, we introduce a novel approach, combining ant colony optimization (ACO) and nave Bayes classifier (NBC) that is, the nave Bayes ant colony optimization (NACO) procedure. We compare NACO with other similar approaches developing a simulation study. We then derive the NACO procedure with the goal to design artificial enzymes with no sequence homology to the extant one. Our final aim is to mimic the natural fold of 200 amino acids 1AGY serine esterase from Fusarium solani.