Search

We consider the protein sequence engineering problem, which aims to find protein sequences with high fitness levels, starting from a given wild-type sequence. Directed evolution has been a dominating paradigm in this field which has an iterative process to generate variants and select via experimental feedback. We demonstrate large language models (LLMs), despite being trained on massive texts, are secretly protein sequence optimizers. With a directed evolutionary method, LLM can perform protein engineering through Pareto and experiment-budget constrained optimization, demonstrating success on both synthetic and experimental fitness landscapes., Comment: Preprint

The annotation (assigning structural chemical identities) of MS/MS spectra remains a significant challenge due to the enormous molecular diversity in biological samples and the limited scope of reference databases. Currently, the vast majority of spectral measurements remain in the "dark chemical space" without structural annotations. To improve annotation, we propose MADGEN (Mass-spec Attends to De Novo Molecular GENeration), a scaffold-based method for de novo molecular structure generation guided by mass spectrometry data. MADGEN operates in two stages: scaffold retrieval and spectra-conditioned molecular generation starting with the scaffold. In the first stage, given an MS/MS spectrum, we formulate scaffold retrieval as a ranking problem and employ contrastive learning to align mass spectra with candidate molecular scaffolds. In the second stage, starting from the retrieved scaffold, we employ the MS/MS spectrum to guide an attention-based generative model to generate the final molecule. Our approach constrains the molecular generation search space, reducing its complexity and improving generation accuracy. We evaluate MADGEN on three datasets (NIST23, CANOPUS, and MassSpecGym) and evaluate MADGEN's performance with a predictive scaffold retriever and with an oracle retriever. We demonstrate the effectiveness of using attention to integrate spectral information throughout the generation process to achieve strong results with the oracle retriever., Comment: preprint

Graph generation is a critical task in numerous domains, including molecular design and social network analysis, due to its ability to model complex relationships and structured data. While most modern graph generative models utilize adjacency matrix representations, this work revisits an alternative approach that represents graphs as sequences of node set and edge set. We advocate for this approach due to its efficient encoding of graphs and propose a novel representation. Based on this representation, we introduce the Graph Generative Pre-trained Transformer (G2PT), an auto-regressive model that learns graph structures via next-token prediction. To further exploit G2PT's capabilities as a general-purpose foundation model, we explore fine-tuning strategies for two downstream applications: goal-oriented generation and graph property prediction. We conduct extensive experiments across multiple datasets. Results indicate that G2PT achieves superior generative performance on both generic graph and molecule datasets. Furthermore, G2PT exhibits strong adaptability and versatility in downstream tasks from molecular design to property prediction., Comment: preprint

Motivation: A major challenge in metabolomics is annotation: assigning molecular structures to mass spectral fragmentation patterns. Despite recent advances in molecule-to-spectra and in spectra-to-molecular fingerprint prediction (FP), annotation rates remain low. Results: We introduce in this paper a novel paradigm (JESTR) for annotation. Unlike prior approaches that explicitly construct molecular fingerprints or spectra, JESTR leverages the insight that molecules and their corresponding spectra are views of the same data and effectively embeds their representations in a joint space. Candidate structures are ranked based on cosine similarity between the embeddings of query spectrum and each candidate. We evaluate JESTR against mol-to-spec and spec-to-FP annotation tools on three datasets. On average, for rank@[1-5], JESTR outperforms other tools by 23.6%-71.6%. We further demonstrate the strong value of regularization with candidate molecules during training, boosting rank@1 performance by 11.4% and enhancing the model's ability to discern between target and candidate molecules. Through JESTR, we offer a novel promising avenue towards accurate annotation, therefore unlocking valuable insights into the metabolome., Comment: 10 pages, 10 figures, 4 tables

The discovery and identification of molecules in biological and environmental samples is crucial for advancing biomedical and chemical sciences. Tandem mass spectrometry (MS/MS) is the leading technique for high-throughput elucidation of molecular structures. However, decoding a molecular structure from its mass spectrum is exceptionally challenging, even when performed by human experts. As a result, the vast majority of acquired MS/MS spectra remain uninterpreted, thereby limiting our understanding of the underlying (bio)chemical processes. Despite decades of progress in machine learning applications for predicting molecular structures from MS/MS spectra, the development of new methods is severely hindered by the lack of standard datasets and evaluation protocols. To address this problem, we propose MassSpecGym -- the first comprehensive benchmark for the discovery and identification of molecules from MS/MS data. Our benchmark comprises the largest publicly available collection of high-quality labeled MS/MS spectra and defines three MS/MS annotation challenges: \textit{de novo} molecular structure generation, molecule retrieval, and spectrum simulation. It includes new evaluation metrics and a generalization-demanding data split, therefore standardizing the MS/MS annotation tasks and rendering the problem accessible to the broad machine learning community. MassSpecGym is publicly available at \url{https://github.com/pluskal-lab/MassSpecGym}.

We present a new method, "kairosis", for aggregating probability forecasts made over a time period of a single outcome determined at the end of that period. Informed by work on Bayesian change-point detection, we begin by constructing for each time during the period a posterior probability that the forecasts before and after this time are distributed differently. The resulting posterior probability mass function is integrated to give a cumulative mass function, which is used to create a weighted median forecast. The effect is to construct an aggregate in which the most heavily weighted forecasts are those which have been made since the probable most recent change in the forecasts' distribution. Kairosis outperforms standard methods, and is especially suitable for geopolitical forecasting tournaments because it is observed to be robust across disparate questions and forecaster distributions., Comment: 26 pages, 8 figures

Advances in generative AI (GenAI) have raised concerns about detecting and discerning AI-generated content from human-generated content. Most existing literature assumes a paradigm where 'expert' organized disinformation creators and flawed AI models deceive 'ordinary' users. Based on longitudinal ethnographic research with misinformation creators and consumers between 2022-2023, we instead find that GenAI supports bricolage work, where non-experts increasingly use GenAI to remix, repackage, and (re)produce content to meet their personal needs and desires. This research yielded four key findings: First, participants primarily used GenAI for creation, rather than truth-seeking. Second, a spreading 'influencer millionaire' narrative drove participants to become content creators, using GenAI as a productivity tool to generate a volume of (often misinformative) content. Third, GenAI lowered the barrier to entry for content creation across modalities, enticing consumers to become creators and significantly increasing existing creators' output. Finally, participants used Gen AI to learn and deploy marketing tactics to expand engagement and monetize their content. We argue for shifting analysis from the public as consumers of AI content to bricoleurs who use GenAI creatively, often without a detailed understanding of its underlying technology. We analyze how these understudied emergent uses of GenAI produce new or accelerated misinformation harms, and their implications for AI products, platforms and policies., Comment: 44 pages, 25 figures

The purpose of this article is to introduce the original results which devoted with the nonlinear control system problems involves of nonlinear differential equations of fractional orders. Thus, this system is described with a mixed of ordinary derivatives in the first and second order that, are unstable before feedback gain. More precisely, we investigate and analysis the nonlinear control system in related to feedback gain matrix. In addition, we prove that the considered system is locally asymptotically stabilizable via certain conditions. Then, this work reinforce through some application examples that programmed for illustrating and showing the stabilizability of the current systems with high efficiency and accuracy., Comment: 9 pages, regular paper

Self-supervised training methods for transformers have demonstrated remarkable performance across various domains. Previous transformer-based models, such as masked autoencoders (MAE), typically utilize a single normalization layer for both the [CLS] symbol and the tokens. We propose in this paper a simple modification that employs separate normalization layers for the tokens and the [CLS] symbol to better capture their distinct characteristics and enhance downstream task performance. Our method aims to alleviate the potential negative effects of using the same normalization statistics for both token types, which may not be optimally aligned with their individual roles. We empirically show that by utilizing a separate normalization layer, the [CLS] embeddings can better encode the global contextual information and are distributed more uniformly in its anisotropic space. When replacing the conventional normalization layer with the two separate layers, we observe an average 2.7% performance improvement over the image, natural language, and graph domains., Comment: NIPS 2023

We conducted ethnographic research with 31 misinformation creators and consumers in Brazil and the US before, during, and after a major election to understand the consumption and production of election and medical misinformation. This study contributes to research on misinformation ecosystems by focusing on poorly understood small players, or "micro-influencers", who create misinformation in peer-to-peer networks. We detail four key tactics that micro-influencers use. First, they typically disseminate "gray area" content rather than expert-falsified claims, using subtle aesthetic and rhetorical tactics to evade moderation. Second, they post in small, closed groups where members feel safe and predisposed to trust content. Third, they explicitly target misinformation consumers' emotional and social needs. Finally, they post a high volume of short, repetitive content to plant seeds of doubt and build trust in influencers as unofficial experts. We discuss the implications these micro-influencers have for misinformation interventions and platforms' efforts to moderate misinformation., Comment: 30 pages, 13 figures, 2 tables

Data provenance, or data lineage, describes the life cycle of data. In scientific workflows on HPC systems, scientists often seek diverse provenance (e.g., origins of data products, usage patterns of datasets). Unfortunately, existing provenance solutions cannot address the challenges due to their incompatible provenance models and/or system implementations. In this paper, we analyze four representative scientific workflows in collaboration with the domain scientists to identify concrete provenance needs. Based on the first-hand analysis, we propose a provenance framework called PROV-IO+, which includes an I/O-centric provenance model for describing scientific data and the associated I/O operations and environments precisely. Moreover, we build a prototype of PROV-IO+ to enable end-to-end provenance support on real HPC systems with little manual effort. The PROV-IO+ framework can support both containerized and non-containerized workflows on different HPC platforms with flexibility in selecting various classes of provenance. Our experiments with realistic workflows show that PROV-IO+ can address the provenance needs of the domain scientists effectively with reasonable performance (e.g., less than 3.5% tracking overhead for most experiments). Moreover, PROV-IO+ outperforms a state-of-the-art system (i.e., ProvLake) in our experiments.

Assessing the trustworthiness of information online is complicated. Literacy-based paradigms are both widely used to help and widely critiqued. We conducted a study with 35 Gen Zers from across the U.S. to understand how they assess information online. We found that they tended to encounter -- rather than search for -- information, and that those encounters were shaped more by social motivations than by truth-seeking queries. For them, information processing is fundamentally a social practice. Gen Zers interpreted online information together, as aspirational members of social groups. Our participants sought information sensibility: a socially-informed awareness of the value of information encountered online. We outline key challenges they faced and practices they used to make sense of information. Our findings suggest that like their information sensibility practices, solutions and strategies to address misinformation should be embedded in social contexts online., Comment: 17 pages, 3 figures, 2 tables, 130 references

While vision transformers have achieved impressive results, effectively and efficiently accelerating these models can further boost performances. In this work, we propose a dense/sparse training framework to obtain a unified model, enabling weight sharing across various token densities. Thus one model offers a range of accuracy and throughput tradeoffs for different applications. Besides, we introduce adaptive token pruning to optimize the patch token sparsity based on the input image. In addition, we investigate knowledge distillation to enhance token selection capability in early transformer modules. Sparse adaptive image Transformer (SaiT) offers varying levels of model acceleration by merely changing the token sparsity on the fly. Specifically, SaiT reduces the computation complexity (FLOPs) by 39% - 43% and increases the throughput by 67% - 91% with less than 0.5% accuracy loss for various vision transformer models. Meanwhile, the same model also provides the zero accuracy drop option by skipping the sparsification step. SaiT achieves better accuracy and computation tradeoffs than state-of-the-art transformer and convolutional models.

Stratifying systems, which have been defined for module, triangulated and exact categories previously, were developed to produce examples of standardly stratified algebras. A stratifying system $\Phi$ is a finite set of objects satisfying some orthogonality conditions. One very interesting property is that the subcategory $\mathcal{F}(\Phi)$ of objects admitting a composition series-like filtration with factors in $\Phi$ has the Jordan-H{\"{o}}lder property on these filtrations. This article has two main aims. First, we introduce notions of subobjects, simple objects and composition series for an extriangulated category, in order to define a Jordan-H{\"{o}}lder extriangulated category. Moreover, we characterise Jordan-H{\"{o}}lder, length, weakly idempotent complete extriangulated categories in terms of the associated Grothendieck monoid and Grothendieck group. Second, we develop a theory of stratifying systems in extriangulated categories. We define projective stratifying systems and show that every stratifying system $\Phi$ in an extriangulated category is part of a minimal projective one $(\Phi,Q)$. We prove that $\mathcal{F}(\Phi)$ is a length, Jordan-H{\"{o}}lder extriangulated category when $(\Phi,Q)$ satisfies a left exactness condition. We give several examples and answer a recent question of Enomoto--Saito in the negative., Comment: v4: 29 pages; updates following referee report, in particular Cor 4.18 improved; other minor changes

The increasing threat of emerging infectious diseases and antimicrobial resistance requires more efficient, high-quality research. Perpetual Observational Studies (POS) nested within a clinical research network can improve planning, quality and efficiency of interventional and observational studies, although real-life benefits and challenges need to be assessed. Ecraid (European Clinical Research Alliance on Infectious Diseases) has initiated POS and will monitor the impact for five specific infectious syndromes., Comment: 16 pages, 2 tables

Though image transformers have shown competitive results with convolutional neural networks in computer vision tasks, lacking inductive biases such as locality still poses problems in terms of model efficiency especially for embedded applications. In this work, we address this issue by introducing attention masks to incorporate spatial locality into self-attention heads. Local dependencies are captured efficiently with masked attention heads along with global dependencies captured by unmasked attention heads. With Masked attention image Transformer - MaiT, top-1 accuracy increases by up to 1.7% compared to CaiT with fewer parameters and FLOPs, and the throughput improves by up to 1.5X compared to Swin. Encoding locality with attention masks is model agnostic, and thus it applies to monolithic, hierarchical, or other novel transformer architectures.

Composite materials achieved by including transition-metal oxides with different structures and morphologies in sulfur are suggested as scalable cathodes for high-energy lithium-sulfur (Li-S) batteries. The composites contain 80 wt.% sulfur and 20 wt.% of either MnO2 or TiO2, leading to a sulfur content in the electrode of 64 wt.% and revealing a reversible, fast, and lowly polarized conversion process in the cell with limited interphase resistance. The S-TiO2 composite exhibits an excellent rate capability between C/10 and 2C, and a cycle life extended over 400 cycles at 2C, owing to the effects of the nanometric TiO2 additive in boosting the reaction kinetics. Instead, the micrometric sized particles of MnO2 partially limit the electrochemical activity of S-MnO2 to the current rate of 1C. Nevertheless, both S-MnO2 and S-TiO2 withstand a sulfur loading up to values approaching 6 mgcm-2, and deliver an areal capacit ranging from about 4.5 to 5.5 mAhcm-2 at C/5. The excellent performances of the metal oxide-sulfur electrodes, even at high active material loading, and the possible scalability of the synthetic pathway adopted in the work suggest that the composites are viable cathodes for next-generation Li-S batteries with high energy density and efficient electrochemical process.

Pruning is an effective way to reduce the huge inference cost of Transformer models. However, prior work on pruning Transformers requires retraining the models. This can add high training cost and high complexity to model deployment, making it difficult to use in many practical situations. To address this, we propose a fast post-training pruning framework for Transformers that does not require any retraining. Given a resource constraint and a sample dataset, our framework automatically prunes the Transformer model using structured sparsity methods. To retain high accuracy without retraining, we introduce three novel techniques: (i) a lightweight mask search algorithm that finds which heads and filters to prune based on the Fisher information; (ii) mask rearrangement that complements the search algorithm; and (iii) mask tuning that reconstructs the output activations for each layer. We apply our method to BERT-base and DistilBERT, and we evaluate its effectiveness on GLUE and SQuAD benchmarks. Our framework achieves up to 2.0x reduction in FLOPs and 1.56x speedup in inference latency, while maintaining < 1% loss in accuracy. Importantly, our framework prunes Transformers in less than 3 minutes on a single GPU, which is over two orders of magnitude faster than existing pruning approaches that retrain the models., Comment: NeurIPS 2022

A key challenge in metabolomics is annotating measured spectra from a biological sample with chemical identities. Currently, only a small fraction of measurements can be assigned identities. Two complementary computational approaches have emerged to address the annotation problem: mapping candidate molecules to spectra, and mapping query spectra to molecular candidates. In essence, the candidate molecule with the spectrum that best explains the query spectrum is recommended as the target molecule. Despite candidate ranking being fundamental in both approaches, no prior works utilized rank learning tasks in determining the target molecule. We propose a novel machine learning model, Ensemble Spectral Prediction (ESP), for metabolite annotation. ESP takes advantage of prior neural network-based annotation models that utilize multilayer perceptron (MLP) networks and Graph Neural Networks (GNNs). Based on the ranking results of the MLP and GNN-based models, ESP learns a weighting for the outputs of MLP and GNN spectral predictors to generate a spectral prediction for a query molecule. Importantly, training data is stratified by molecular formula to provide candidate sets during model training. Further, baseline MLP and GNN models are enhanced by considering peak dependencies through multi-head attention mechanism and multi-tasking on spectral topic distributions. ESP improves average rank by 41% and 30% over the MLP and GNN baselines, respectively, demonstrating remarkable performance gain over state-of-the-art neural network approaches. We show that annotation performance, for ESP and other models, is a strong function of the number of molecules in the candidate set and their similarity to the target molecule.

A LiFe0.6Mn0.4PO4 (LFMP) cathode exploiting the olivine structure is herein synthesized and characterized in terms of structure, morphology, and electrochemical features in a lithium cell. The material shows reversibility of the electrochemical process which evolves at 3.5 and 4 V versus Li+/Li due to the Fe+2/Fe+3 and Mn+2/Mn+3 redox couples, respectively, as determined by cyclic voltammetry. The LFMP has a well-defined olivine structure revealed by X-ray diffraction, a morphology consisting of submicron particle aggregated into micrometric clusters as indicated by scanning and transmission electron microscopy, with a carbon weight ratio of about 4% as suggested by thermogravimetry. The electrode is used in lithium cells subjected to galvanostatic cycling with a conventional liquid electrolyte, and demonstrates a maximum capacity of 130 mAh g-1, satisfactory rate capability, excellent efficiency, and a stable trend. Therefore, the material is studied in a lithium metal polymer cell exploiting an electrolyte based on polyethylene glycol dimethyl ether with a solid configuration. The cell reveals very promising features in terms of capacity, efficiency, and retention, and suggests the LFMP material as a suitable electrode for polymer batteries characterized by increased energy density and remarkable safety.

The concept of green in a battery involves the chemical nature of electrodes and electrolytes as well as the economic sustainability of the cell. Although these aspects are typically discussed separately, they are deeply interconnected: indeed, a new electrolyte can allow the use of different cathodes with higher energy, lower cost or more pronounced environmental compatibility. We focus on alternative class of electrolyte solutions for lithium batteries formed by dissolving LiX salts in glyme solvents, i.e., organic ethers with the molecular formula CH3O[CH2CH2O]nCH3 differing by chain length. The advantages of these electrolytes are illustrated in terms of flammability, stability, toxicity, environmental compatibility, cell performances and economic impact. A particular light is shed on the stability of these systems, particularly in the polymer state, and in various environments including oxygen, sulfur and high-energy lithium metal. The most relevant studies on the chemical-physical features, the characteristic structures, the favorable properties, and the electrochemical behavior of the glyme-based solutions are discussed, and the most recent technological achievements in terms of cell design and battery performance are described. The use of glyme-based electrolytes in high-energy cells arranged by coupling the lithium-metal anode with conventional insertion cathodes as well as in alternative and new batteries exploiting the Li-S and Li-O2 conversion processes are described in detail. The paragraphs reveal bonuses, including safety, low cost and sustainability, that can be achieved by employing the glyme-based electrolytes with respect to the commercially available ones, in particular taking into account future and alternative applications. Particular relevance is given by the glymes with long chain that reveal a remarkable stability, high safety and very low toxicity.

Accurately predicting the likelihood of interaction between two objects (compound-protein sequence, user-item, author-paper, etc.) is a fundamental problem in Computer Science. Current deep-learning models rely on learning accurate representations of the interacting objects. Importantly, relationships between the interacting objects, or features of the interaction, offer an opportunity to partition the data to create multi-views of the interacting objects. The resulting congruent and non-congruent views can then be exploited via contrastive learning techniques to learn enhanced representations of the objects., Comment: 11 pages, submitted to BioInformatics

A safe lithium-sulfur (Li-S) battery employs a composite polymer electrolyte based on a poly(ethylene glycol) dimethyl ether (PEGDME) solid at room temperature. The electrolyte membrane enables a stable and reversible Li-S electrochemical process already at 50{\deg}C, with low resistance at the electrode/electrolyte interphase and fast Li+ transport. The relatively low molecular weight of the PEGDME and the optimal membrane composition in terms of salts and ceramic allow a liquid-like Li-S conversion reaction by heating at moderately high temperature, still holding the solid-like polymer state of the cell. Therefore, the electrochemical reaction of the polymer Li-S cell is characterized by the typical dissolution of lithium polysulfides into the electrolyte medium during discharge and the subsequent deposition of sulfur at the electrode/electrolyte interphase during charge. On the other hand, the remarkable thermal stability of the composite polymer electrolyte (up to 300{\deg}C) suggests a lithium-metal battery with safety content significantly higher than that using the common, flammable liquid solutions. Hence, the Li-S polymer battery delivers at 50{\deg}C and 2 V a stable capacity approaching 700 mAhgS-1, with a steady-state coulombic efficiency of 98%. These results suggest a novel, alternative approach to achieve safe, high energy batteries with solid polymer configuration.

Motivated by problems in the neural networks setting, we study moduli spaces of double framed quiver representations and give both a linear algebra description and a representation theoretic description of these moduli spaces. We define a network category whose isomorphism classes of objects correspond to the orbits of quiver representations, in which neural networks map input data. We then prove that the output of a neural network depends only on the corresponding point in the moduli space. Finally, we present a different perspective on mapping neural networks with a specific activation function, called ReLU, to a moduli space using the symplectic reduction approach to quiver moduli., Comment: 27 pages

Despite experimental and curation efforts, the extent of enzyme promiscuity on substrates continues to be largely unexplored and under documented. Recommender systems (RS), which are currently unexplored for the enzyme-substrate interaction prediction problem, can be utilized to provide enzyme recommendations for substrates, and vice versa. The performance of Collaborative-Filtering (CF) recommender systems however hinges on the quality of embedding vectors of users and items (enzymes and substrates in our case). Importantly, enhancing CF embeddings with heterogeneous auxiliary data, specially relational data (e.g., hierarchical, pairwise, or groupings), remains a challenge. We propose an innovative general RS framework, termed Boost-RS, that enhances RS performance by "boosting" embedding vectors through auxiliary data. Specifically, Boost-RS is trained and dynamically tuned on multiple relevant auxiliary learning tasks Boost-RS utilizes contrastive learning tasks to exploit relational data. To show the efficacy of Boost-RS for the enzyme-substrate prediction interaction problem, we apply the Boost-RS framework to several baseline CF models. We show that each of our auxiliary tasks boosts learning of the embedding vectors, and that contrastive learning using Boost-RS outperforms attribute concatenation and multi-label learning. We also show that Boost-RS outperforms similarity-based models. Ablation studies and visualization of learned representations highlight the importance of using contrastive learning on some of the auxiliary data in boosting the embedding vectors., Comment: 9 pages; 2 figures

The degradation mechanism in a sodium cell of a layered Na0.48Al0.03Co0.18Ni0.18Mn0.47O2 (NCAM) cathode with P3/P2 structure is investigated by revealing the changes in microstructure and composition upon cycling. The work aims to rationalize the gradual performance decay and the alteration of the electrochemical response in terms of polarization, voltage signature, and capacity loss. Spatial reconstructions of the electrode by X-ray computed tomography at the nanoscale supported by quantitative and qualitative analyses show fractures and deformations in the cycled layered metal-oxide particles, as well as inorganic side compounds deposited on the material. These irreversible morphological modifications reflect structural heterogeneities across the cathode particles due to formation of various domains with different Na+ intercalation degrees. Besides, X-ray photoelectron spectroscopy data suggest that the latter inorganic species in the cycled electrode are mainly composed of NaF, Na2O, and NaCO3 formed by parasitic electrolyte decomposition. The precipitation of these insulating compounds at the electrode/electrolyte interphase and the related structural stresses induced in the material lead to a decrease in cathode particle size and partial loss of electrochemical activity. The retention of the NCAM phase after cycling suggests that electrolyte upgrade may improve the performance of the cathode to achieve practical application for sustainable energy storage.

A Li-conversion alpha-Fe2O3@C nanocomposite anode and a high-voltage LiNi0.5Mn1.5O4 cathode are synthesized in parallel, characterized, and combined in a Li-ion battery. alpha-Fe2O3@C is prepared via annealing of maghemite iron oxide and sucrose under an argon atmosphere and subsequent oxidation in air. The nanocomposite exhibits a satisfactory electrochemical response in a lithium half-cell, delivering almost 900 mA h g-1, as well as a significantly longer cycle life and higher rate capability compared to the bare iron oxide precursor. The LiNi0.5Mn1.5O4 cathode, achieved using a modified co-precipitation approach, reveals a well-defined spinel structure without impurities, a sub-micrometrical morphology, and a reversible capacity of ca. 120 mA h g-1 in a lithium half-cell with an operating voltage of 4.8 V. Hence, a lithium-ion battery is assembled by coupling the alpha-Fe2O3@C anode with the LiNi0.5Mn1.5O4 cathode. This cell operates at about 3.2 V, delivering a stable capacity of 110 mA h g-1 (referred to the cathode mass) with a Coulombic efficiency exceeding 97%. Therefore, this cell is suggested as a promising energy storage system with expected low economic and environmental impacts.