A review of active learning approaches to experimental design for uncovering biological networks

Current and future directions in network biology

Summary

Network biology is an interdisciplinary field bridging computational and biological sciences that has proved pivotal in advancing the understanding of cellular functions and diseases across biological systems and scales. Although the field has been around for two decades, it remains nascent. It has witnessed rapid evolution, accompanied by emerging challenges. These stem from various factors, notably the growing complexity and volume of data together with the increased diversity of data types describing different tiers of biological organization. We discuss prevailing research directions in network biology, focusing on molecular/cellular networks but also on other biological network types such as biomedical knowledge graphs, patient similarity networks, brain networks, and social/contact networks relevant to disease spread. In more detail, we highlight areas of inference and comparison of biological networks, multimodal data integration and heterogeneous networks, higher-order network analysis, machine learning on networks, and network-based personalized medicine. Following the overview of recent breakthroughs across these five areas, we offer a perspective on future directions of network biology. Additionally, we discuss scientific communities, educational initiatives, and the importance of fostering diversity within the field. This article establishes a roadmap for an immediate and long-term vision for network biology.

Availability and implementation

Not applicable.

Technologies for whole-cell modeling: Genome-wide reconstruction of a cell in silico

With advances in high-throughput, large-scale in vivo measurement and genome modification techniques at the single-nucleotide level, there is an increasing demand for the development of new technologies for the flexible design and control of cellular systems. Computer-aided design is a powerful tool to design new cells. Whole-cell modeling aims to integrate various cellular subsystems, determine their interactions and cooperative mechanisms, and predict comprehensive cellular behaviors by computational simulations on a genome-wide scale. It has been applied to prokaryotes, yeasts, and higher eukaryotic cells, and utilized in a wide range of applications, including production of valuable substances, drug discovery, and controlled differentiation. Whole-cell modeling, consisting of several thousand elements with diverse scales and properties, requires innovative model construction, simulation, and analysis techniques. Furthermore, whole-cell modeling has been extended to multiple scales, including high-resolution modeling at the single-nucleotide and single-amino acid levels and multicellular modeling of tissues and organs. This review presents an overview of the current state of whole-cell modeling, discusses the novel computational and experimental technologies driving it, and introduces further developments toward multihierarchical modeling on a whole-genome scale.

RECOVER identifies synergistic drug combinations in vitro through sequential model optimization

For large libraries of small molecules, exhaustive combinatorial chemical screens become infeasible to perform when considering a range of disease models, assay conditions, and dose ranges. Deep learning models have achieved state-of-the-art results in silico for the prediction of synergy scores. However, databases of drug combinations are biased toward synergistic agents and results do not generalize out of distribution. During 5 rounds of experimentation, we employ sequential model optimization with a deep learning model to select drug combinations increasingly enriched for synergism and active against a cancer cell line-evaluating only ∼5% of the total search space. Moreover, we find that learned drug embeddings (using structural information) begin to reflect biological mechanisms. In silico benchmarking suggests search queries are ∼5-10× enriched for highly synergistic drug combinations by using sequential rounds of evaluation when compared with random selection or ∼3× when using a pretrained model.

Active learning applied to automated physical systems increases the rate of discovery

Active machine learning is widely used in computational studies where repeated numerical simulations can be conducted on high performance computers without human intervention. But translation of these active learning methods to physical systems has proven more difficult and the accelerated pace of discoveries aided by these methods remains as yet unrealized. Through the presentation of a general active learning framework and its application to large-scale boundary layer wind tunnel experiments, we demonstrate that the active learning framework used so successfully in computational studies is directly applicable to the investigation of physical experimental systems and the corresponding improvements in the rate of discovery can be transformative. We specifically show that, for our wind tunnel experiments, we are able to achieve in approximately 300 experiments a learning objective that would be impossible using traditional methods.

Improving Active Learning Performance through the Use of Data Augmentation

Active learning (AL) is a well-known technique to optimize data usage in training, through the interactive selection of unlabeled observations, out of a large pool of unlabeled data, to be labeled by a supervisor. Its focus is to find the unlabeled observations that, once labeled, will maximize the informativeness of the training dataset, therefore reducing data-related costs. The literature describes several methods to improve the effectiveness of this process. Nonetheless, there is a paucity of research developed around the application of artificial data sources in AL, especially outside image classification or NLP. This paper proposes a new AL framework, which relies on the effective use of artificial data. It may be used with any classifier, generation mechanism, and data type and can be integrated with multiple other state-of-the-art AL contributions. This combination is expected to increase the ML classifier’s performance and reduce both the supervisor’s involvement and the amount of required labeled data at the expense of a marginal increase in computational time. The proposed method introduces a hyperparameter optimization component to improve the generation of artificial instances during the AL process as well as an uncertainty-based data generation mechanism. We compare the proposed method to the standard framework and an oversampling-based active learning method for more informed data generation in an AL context. The models’ performance was tested using four different classifiers, two AL-specific performance metrics, and three classification performance metrics over 15 different datasets. We demonstrated that the proposed framework, using data augmentation, significantly improved the performance of AL, both in terms of classification performance and data selection efficiency (all the codes and preprocessed data developed for this study are available at https://github.com/joaopfonseca/publications/).

High-throughput biochemistry in RNA sequence space: predicting structure and function

RNAs are central to fundamental biological processes in all known organisms. The set of possible intramolecular interactions of RNA nucleotides defines the range of alternative structural conformations of a specific RNA that can coexist, and these structures enable functional catalytic properties of RNAs and/or their productive intermolecular interactions with other RNAs or proteins. However, the immense combinatorial space of potential RNA sequences has precluded predictive mapping between RNA sequence and molecular structure and function. Recent advances in high-throughput approaches in vitro have enabled quantitative thermodynamic and kinetic measurements of RNA-RNA and RNA-protein interactions, across hundreds of thousands of sequence variations. In this Review, we explore these techniques, how they can be used to understand RNA function and how they might form the foundations of an accurate model to predict the structure and function of an RNA directly from its nucleotide sequence. The experimental techniques and modelling frameworks discussed here are also highly relevant for the sampling of sequence-structure-function space of DNAs and proteins.

Anti-tumor effects of cryptotanshinone (C19H20O3) in human osteosarcoma cell lines

Osteosarcoma is the most prevalent malignant bone tumor and occurs most commonly in the adolescent and young adult population. Despite the recent advances in surgeries and chemotherapy, the overall survival in patients with resectable metastases is around 20%. This challenge in osteosarcoma is often attributed to the drastic differences in the tumorigenic profiles and mutations among patients. With diverse mutations and multiple oncogenes, it is necessary to identify the therapies that can attack various mutations and simultaneously have minor side-effects. In this paper, we constructed the osteosarcoma pathway from literature and modeled it using ordinary differential equations. We then simulated this network for every possible gene mutation and their combinations and ranked different drug combinations based on their efficacy to drive a mutated osteosarcoma network towards cell death. Our theoretical results predict that drug combinations with Cryptotanshinone (C₁₉H₂₀O₃), a traditional Chinese herb derivative, have the best overall performance. Specifically, Cryptotanshinone in combination with Temsirolimus inhibit the JAK/STAT, MAPK/ERK, and PI3K/Akt/mTOR pathways and induce cell death in tumor cells. We corroborated our theoretical predictions using wet-lab experiments on SaOS2, 143B, G292, and HU03N1 human osteosarcoma cell lines, thereby demonstrating the potency of Cryptotanshinone in fighting osteosarcoma.

Identification of dynamic driver sets controlling phenotypical landscapes

Controlling phenotypical landscapes is of vital interest to modern biology. This task becomes highly demanding because cellular decisions involve complex networks engaging in crosstalk interactions. Previous work on control theory indicates that small sets of compounds can control single phenotypes. However, a dynamic approach is missing to determine the drivers of the whole network dynamics. By analyzing 35 biologically motivated Boolean networks, we developed a method to identify small sets of compounds sufficient to decide on the entire phenotypical landscape. These compounds do not strictly prefer highly related compounds and show a smaller impact on the stability of the attractor landscape. The dynamic driver sets include many intervention targets and cellular reprogramming drivers in human networks. Finally, by using a new comprehensive model of colorectal cancer, we provide a complete workflow on how to implement our approach to shift from in silico to in vitro guided experiments.

Utilizing graph machine learning within drug discovery and development

Graph machine learning (GML) is receiving growing interest within the pharmaceutical and biotechnology industries for its ability to model biomolecular structures, the functional relationships between them, and integrate multi-omic datasets - amongst other data types. Herein, we present a multidisciplinary academic-industrial review of the topic within the context of drug discovery and development. After introducing key terms and modelling approaches, we move chronologically through the drug development pipeline to identify and summarize work incorporating: target identification, design of small molecules and biologics, and drug repurposing. Whilst the field is still emerging, key milestones including repurposed drugs entering in vivo studies, suggest GML will become a modelling framework of choice within biomedical machine learning.

Categorical Matrix Completion With Active Learning for High-Throughput Screening

The recent advances in wet-lab automation enable high-throughput experiments to be conducted seamlessly. In particular, the exhaustive enumeration of all possible conditions is always involved in high-throughput screening. Nonetheless, such a screening strategy is hardly believed to be optimal and cost-effective. By incorporating artificial intelligence, we design an open-source model based on categorical matrix completion and active machine learning to guide high throughput screening experiments. Specifically, we narrow our scope to the high-throughput screening for chemical compound effects on diverse protein sub-cellular locations. In the proposed model, we believe that exploration is more important than the exploitation in the long-run of high-throughput screening experiment, Therefore, we design several innovations to circumvent the existing limitations. In particular, categorical matrix completion is designed to accurately impute the missing experiments while margin sampling is also implemented for uncertainty estimation. The model is systematically tested on both simulated and real data. The simulation results reflect that our model can be robust to diverse scenarios, while the real data results demonstrate the wet-lab applicability of our model for high-throughput screening experiments. Lastly, we attribute the model success to its exploration ability by revealing the related matrix ranks and distinct experiment coverage comparisons.

On the hierarchical design of biochemical-based digital computations

The understanding of the biochemical processes underpinning various biological systems has significantly increased in recent decades and has even prompted reverse engineering of certain of life's more complex processes. The most prominent example is modern computers designed to mimic neuron activity. This work forms part of growing endeavors to return advances in the theory of computation and electronics to biology. In this context, we present a set of requirements sufficient for the design of biochemical analogs of modern electronics in a hierarchical, modular fashion that mimics the design of modern computational devices. This theoretical approach is based on a simple enzymatic analog of the transistor and supported by numerical simulations of biochemical models of enzymatic networks equivalent to complex, and modular, interconnecting electronic circuitry (including clocks, Flip-Flops, adders, decoders, and multiplexers). Furthermore, the proposed approach has been implemented in the form of a Python library capable of creating and testing models of complex bio-analog digital computations based on the execution of an elementary universal logic gate. In tribute to Claude Shannon, our biochemical network materializes his example of a "password" recognition that moves the language of the modern theory of automata beyond combinatorial logic and towards sequential logic.

Regression plane concept for analysing continuous cellular processes with machine learning

Biological processes are inherently continuous, and the chance of phenotypic discovery is significantly restricted by discretising them. Using multi-parametric active regression we introduce the Regression Plane (RP), a user-friendly discovery tool enabling class-free phenotypic supervised machine learning, to describe and explore biological data in a continuous manner. First, we compare traditional classification with regression in a simulated experimental setup. Second, we use our framework to identify genes involved in regulating triglyceride levels in human cells. Subsequently, we analyse a time-lapse dataset on mitosis to demonstrate that the proposed methodology is capable of modelling complex processes at infinite resolution. Finally, we show that hemocyte differentiation in Drosophila melanogaster has continuous characteristics.

Network analyses in microbiome based on high-throughput multi-omics data

Together with various hosts and environments, ubiquitous microbes interact closely with each other forming an intertwined system or community. Of interest, shifts of the relationships between microbes and their hosts or environments are associated with critical diseases and ecological changes. While advances in high-throughput Omics technologies offer a great opportunity for understanding the structures and functions of microbiome, it is still challenging to analyse and interpret the omics data. Specifically, the heterogeneity and diversity of microbial communities, compounded with the large size of the datasets, impose a tremendous challenge to mechanistically elucidate the complex communities. Fortunately, network analyses provide an efficient way to tackle this problem, and several network approaches have been proposed to improve this understanding recently. Here, we systemically illustrate these network theories that have been used in biological and biomedical research. Then, we review existing network modelling methods of microbial studies at multiple layers from metagenomics to metabolomics and further to multi-omics. Lastly, we discuss the limitations of present studies and provide a perspective for further directions in support of the understanding of microbial communities.

Machine-learning model led design to experimentally test species thermal limits: The case of kissing bugs (Triatominae)

Species Distribution Modelling (SDM) determines habitat suitability of a species across geographic areas using macro-climatic variables; however, micro-habitats can buffer or exacerbate the influence of macro-climatic variables, requiring links between physiology and species persistence. Experimental approaches linking species physiology to micro-climate are complex, time consuming and expensive. E.g., what combination of exposure time and temperature is important for a species thermal tolerance is difficult to judge a priori. We tackled this problem using an active learning approach that utilized machine learning methods to guide thermal tolerance experimental design for three kissing-bug species: Triatoma infestans, Rhodnius prolixus, and Panstrongylus megistus (Hemiptera: Reduviidae: Triatominae), vectors of the parasite causing Chagas disease. As with other pathogen vectors, triatomines are well known to utilize micro-habitats and the associated shift in microclimate to enhance survival. Using a limited literature-collected dataset, our approach showed that temperature followed by exposure time were the strongest predictors of mortality; species played a minor role, and life stage was the least important. Further, we identified complex but biologically plausible nonlinear interactions between temperature and exposure time in shaping mortality, together setting the potential thermal limits of triatomines. The results from this data led to the design of new experiments with laboratory results that produced novel insights of the effects of temperature and exposure for the triatomines. These results, in turn, can be used to better model micro-climatic envelope for the species. Here we demonstrate the power of an active learning approach to explore experimental space to design laboratory studies testing species thermal limits. Our analytical pipeline can be easily adapted to other systems and we provide code to allow practitioners to perform similar analyses. Not only does our approach have the potential to save time and money: it can also increase our understanding of the links between species physiology and climate, a topic of increasing ecological importance.

Use of artificial intelligence to enhance phenotypic drug discovery

Research and development (R&D) productivity across the pharmaceutical industry has received close scrutiny over the past two decades, especially taking into consideration reports of attrition rates and the colossal cost for drug development. The respective merits of the two main drug discovery approaches, phenotypic and target based, have divided opinion across the research community, because each hold different advantages for identifying novel molecular entities with a successful path to the market. Nevertheless, both have low translatability in the clinic. Artificial intelligence (AI) and adoption of machine learning (ML) tools offer the promise of revolutionising drug development, and overcoming obstacles in the drug discovery pipeline. Here, we assess the potential of target-driven and phenotypic-based approaches and offer a holistic description of the current state of the field, from both a scientific and industry perspective. With the emerging partnerships between AI/ML and pharma still in their relative infancy, we investigate the potential and current limitations with a particular focus on phenotypic drug discovery. Finally, we emphasise the value of public-private partnerships (PPPs) and cross-disciplinary collaborations to foster innovation and facilitate efficient drug discovery programmes.

Optimal experimental design for mathematical models of haematopoiesis

The haematopoietic system has a highly regulated and complex structure in which cells are organized to successfully create and maintain new blood cells. It is known that feedback regulation is crucial to tightly control this system, but the specific mechanisms by which control is exerted are not completely understood. In this work, we aim to uncover the underlying mechanisms in haematopoiesis by conducting perturbation experiments, where animal subjects are exposed to an external agent in order to observe the system response and evolution. We have developed a novel Bayesian hierarchical framework for optimal design of perturbation experiments and proper analysis of the data collected. We use a deterministic model that accounts for feedback and feedforward regulation on cell division rates and self-renewal probabilities. A significant obstacle is that the experimental data are not longitudinal, rather each data point corresponds to a different animal. We overcome this difficulty by modelling the unobserved cellular levels as latent variables. We then use principles of Bayesian experimental design to optimally distribute time points at which the haematopoietic cells are quantified. We evaluate our approach using synthetic and real experimental data and show that an optimal design can lead to better estimates of model parameters.

Deep Learning for Protein-Protein Interaction Site Prediction

Protein-protein interactions (PPIs) are central to cellular functions. Experimental methods for predicting PPIs are well developed but are time and resource expensive and suffer from high false-positive error rates at scale. Computational prediction of PPIs is highly desirable for a mechanistic understanding of cellular processes and offers the potential to identify highly selective drug targets. In this chapter, details of developing a deep learning approach to predicting which residues in a protein are involved in forming a PPI-a task known as PPI site prediction-are outlined. The key decisions to be made in defining a supervised machine learning project in this domain are here highlighted. Alternative training regimes for deep learning models to address shortcomings in existing approaches and provide starting points for further research are discussed. This chapter is written to serve as a companion to developing deep learning approaches to protein-protein interaction site prediction, and an introduction to developing geometric deep learning projects operating on protein structure graphs.

Leveraging Uncertainty in Machine Learning Accelerates Biological Discovery and Design

Machine learning that generates biological hypotheses has transformative potential, but most learning algorithms are susceptible to pathological failure when exploring regimes beyond the training data distribution. A solution to address this issue is to quantify prediction uncertainty so that algorithms can gracefully handle novel phenomena that confound standard methods. Here, we demonstrate the broad utility of robust uncertainty prediction in biological discovery. By leveraging Gaussian process-based uncertainty prediction on modern pre-trained features, we train a model on just 72 compounds to make predictions over a 10,833-compound library, identifying and experimentally validating compounds with nanomolar affinity for diverse kinases and whole-cell growth inhibition of Mycobacterium tuberculosis. Uncertainty facilitates a tight iterative loop between computation and experimentation and generalizes across biological domains as diverse as protein engineering and single-cell transcriptomics. More broadly, our work demonstrates that uncertainty should play a key role in the increasing adoption of machine learning algorithms into the experimental lifecycle.

Accelerated knowledge discovery from omics data by optimal experimental design

How to design experiments that accelerate knowledge discovery on complex biological landscapes remains a tantalizing question. We present an optimal experimental design method (coined OPEX) to identify informative omics experiments using machine learning models for both experimental space exploration and model training. OPEX-guided exploration of Escherichia coli’s populations exposed to biocide and antibiotic combinations lead to more accurate predictive models of gene expression with 44% less data. Analysis of the proposed experiments shows that broad exploration of the experimental space followed by fine-tuning emerges as the optimal strategy. Additionally, analysis of the experimental data reveals 29 cases of cross-stress protection and 4 cases of cross-stress vulnerability. Further validation reveals the central role of chaperones, stress response proteins and transport pumps in cross-stress exposure. This work demonstrates how active learning can be used to guide omics data collection for training predictive models, making evidence-driven decisions and accelerating knowledge discovery in life sciences.

How to design experiments that accelerate knowledge discovery on complex biological landscapes remains a tantalizing question. Here, the authors present OPEX, an optimal experimental design method to identify informative omics experiments for both experimental space exploration and model training.

Retrieving Quantum Information with Active Learning

Active learning is a machine learning method aiming at optimal design for model training. At variance with supervised learning, which labels all samples, active learning provides an improved model by labeling samples with maximal uncertainty according to the estimation model. Here, we propose the use of active learning for efficient quantum information retrieval, which is a crucial task in the design of quantum experiments. Meanwhile, when dealing with large data output, we employ active learning for the sake of classification with minimal cost in fidelity loss. Indeed, labeling only 5% samples, we achieve almost 90% rate estimation. The introduction of active learning methods in the data analysis of quantum experiments will enhance applications of quantum technologies.

The single-cell eQTLGen consortium

In recent years, functional genomics approaches combining genetic information with bulk RNA-sequencing data have identified the downstream expression effects of disease-associated genetic risk factors through so-called expression quantitative trait locus (eQTL) analysis. Single-cell RNA-sequencing creates enormous opportunities for mapping eQTLs across different cell types and in dynamic processes, many of which are obscured when using bulk methods. Rapid increase in throughput and reduction in cost per cell now allow this technology to be applied to large-scale population genetics studies. To fully leverage these emerging data resources, we have founded the single-cell eQTLGen consortium (sc-eQTLGen), aimed at pinpointing the cellular contexts in which disease-causing genetic variants affect gene expression. Here, we outline the goals, approach and potential utility of the sc-eQTLGen consortium. We also provide a set of study design considerations for future single-cell eQTL studies.

Data science for the scientific life cycle

Data science can be incorporated into every stage of a scientific study. Here we describe how data science can be used to generate hypotheses, to design experiments, to perform experiments, and to analyse data. We also present our vision for how data science techniques will be an integral part of the laboratory of the future.

Computational prediction of gene regulatory networks in plant growth and development

Plants integrate a wide range of cellular, developmental, and environmental signals to regulate complex patterns of gene expression. Recent advances in genomic technologies enable differential gene expression analysis at a systems level, allowing for improved inference of the network of regulatory interactions between genes. These gene regulatory networks, or GRNs, are used to visualize the causal regulatory relationships between regulators and their downstream target genes. Accordingly, these GRNs can represent spatial, temporal, and/or environmental regulations and can identify functional genes. This review summarizes recent computational approaches applied to different types of gene expression data to infer GRNs in the context of plant growth and development. Three stages of GRN inference are described: first, data collection and analysis based on the dataset type; second, network inference application based on data availability and proposed hypotheses; and third, validation based on in silico, in vivo, and in planta methods. In addition, this review relates data collection strategies to biological questions, organizes inference algorithms based on statistical methods and data types, discusses experimental design considerations, and provides guidelines for GRN inference with an emphasis on the benefits of integrative approaches, especially when a priori information is limited. Finally, this review concludes that computational frameworks integrating large-scale heterogeneous datasets are needed for a more accurate (e.g. fewer false interactions), detailed (e.g. discrimination between direct versus indirect interactions), and comprehensive (e.g. genetic regulation under various conditions and spatial locations) inference of GRNs.

New Drosophila Long-Term Memory Genes Revealed by Assessing Computational Function Prediction Methods

A major bottleneck to our understanding of the genetic and molecular foundation of life lies in the ability to assign function to a gene and, subsequently, a protein. Traditional molecular and genetic experiments can provide the most reliable forms of identification, but are generally low-throughput, making such discovery and assignment a daunting task. The bottleneck has led to an increasing role for computational approaches. The Critical Assessment of Functional Annotation (CAFA) effort seeks to measure the performance of computational methods. In CAFA3, we performed selected screens, including an effort focused on long-term memory. We used homology and previous CAFA predictions to identify 29 key Drosophila genes, which we tested via a long-term memory screen. We identify 11 novel genes that are involved in long-term memory formation and show a high level of connectivity with previously identified learning and memory genes. Our study provides first higher-order behavioral assay and organism screen used for CAFA assessments and revealed previously uncharacterized roles of multiple genes as possible regulators of neuronal plasticity at the boundary of information acquisition and memory formation.

An experimental design framework for Markovian gene regulatory networks under stationary control policy

BACKGROUND

A fundamental problem for translational genomics is to find optimal therapies based on gene regulatory intervention. Dynamic intervention involves a control policy that optimally reduces a cost function based on phenotype by externally altering the state of the network over time. When a gene regulatory network (GRN) model is fully known, the problem is addressed using classical dynamic programming based on the Markov chain associated with the network. When the network is uncertain, a Bayesian framework can be applied, where policy optimality is with respect to both the dynamical objective and the uncertainty, as characterized by a prior distribution. In the presence of uncertainty, it is of great practical interest to develop an experimental design strategy and thereby select experiments that optimally reduce a measure of uncertainty.

RESULTS

In this paper, we employ mean objective cost of uncertainty (MOCU), which quantifies uncertainty based on the degree to which uncertainty degrades the operational objective, that being the cost owing to undesirable phenotypes. We assume that a number of conditional probabilities characterizing regulatory relationships among genes are unknown in the Markovian GRN. In sum, there is a prior distribution which can be updated to a posterior distribution by observing a regulatory trajectory, and an optimal control policy, known as an "intrinsically Bayesian robust" (IBR) policy. To obtain a better IBR policy, we select an experiment that minimizes the MOCU remaining after applying its output to the network. At this point, we can either stop and find the resulting IBR policy or proceed to determine more unknown conditional probabilities via regulatory observation and find the IBR policy from the resulting posterior distribution. For sequential experimental design this entire process is iterated. Owing to the computational complexity of experimental design, which requires computation of many potential IBR policies, we implement an approximate method utilizing mean first passage times (MFPTs) - but only in experimental design, the final policy being an IBR policy.

CONCLUSIONS

Comprehensive performance analysis based on extensive simulations on synthetic and real GRNs demonstrate the efficacy of the proposed method, including the accuracy and computational advantage of the approximate MFPT-based design.

Experimental search for high-temperature ferroelectric perovskites guided by two-step machine learning

Experimental search for high-temperature ferroelectric perovskites is a challenging task due to the vast chemical space and lack of predictive guidelines. Here, we demonstrate a two-step machine learning approach to guide experiments in search of xBi[Formula: see text]O3-(1 - x)PbTiO3-based perovskites with high ferroelectric Curie temperature. These involve classification learning to screen for compositions in the perovskite structures, and regression coupled to active learning to identify promising perovskites for synthesis and feedback. The problem is challenging because the search space is vast, spanning ~61,500 compositions and only 167 are experimentally studied. Furthermore, not every composition can be synthesized in the perovskite phase. In this work, we predict x, y, Me', and Me″ such that the resulting compositions have both high Curie temperature and form in the perovskite structure. Outcomes from both successful and failed experiments then iteratively refine the machine learning models via an active learning loop. Our approach finds six perovskites out of ten compositions synthesized, including three previously unexplored {Me'Me″} pairs, with 0.2Bi(Fe0.12Co0.88)O3-0.8PbTiO3 showing the highest measured Curie temperature of 898 K among them.

Data integration and predictive modeling methods for multi-omics datasets

We provide an overview of opportunities and challenges in multi-omics predictive analytics with particular emphasis on data integration and machine learning methods.

Systematic Identification of Machine-Learning Models Aimed to Classify Critical Residues for Protein Function from Protein Structure

Protein structure and protein function should be related, yet the nature of this relationship remains unsolved. Mapping the critical residues for protein function with protein structure features represents an opportunity to explore this relationship, yet two important limitations have precluded a proper analysis of the structure-function relationship of proteins: (i) the lack of a formal definition of what critical residues are and (ii) the lack of a systematic evaluation of methods and protein structure features. To address this problem, here we introduce an index to quantify the protein-function criticality of a residue based on experimental data and a strategy aimed to optimize both, descriptors of protein structure (physicochemical and centrality descriptors) and machine learning algorithms, to minimize the error in the classification of critical residues. We observed that both physicochemical and centrality descriptors of residues effectively relate protein structure and protein function, and that physicochemical descriptors better describe critical residues. We also show that critical residues are better classified when residue criticality is considered as a binary attribute (i.e., residues are considered critical or not critical). Using this binary annotation for critical residues 8 models rendered accurate and non-overlapping classification of critical residues, confirming the multi-factorial character of the structure-function relationship of proteins.