A systematic evaluation of deep learning methods for the prediction of drug synergy in cancer.

One of the main obstacles to the successful treatment of cancer is the phenomenon of drug resistance. A common strategy to overcome resistance is the use of combination therapies. However, the space of possibilities is huge and efficient search strategies are required. Machine Learning (ML) can be a useful tool for the discovery of novel, clinically relevant anti-cancer drug combinations. In particular, deep learning (DL) has become a popular choice for modeling drug combination effects. Here, we set out to examine the impact of different methodological choices on the performance of multimodal DL-based drug synergy prediction methods, including the use of different input data types, preprocessing steps and model architectures. Focusing on the NCI ALMANAC dataset, we found that feature selection based on prior biological knowledge has a positive impact-limiting gene expression data to cancer or drug response-specific genes improved performance. Drug features appeared to be more predictive of drug response, with a 41% increase in coefficient of determination (R2) and 26% increase in Spearman correlation relative to a baseline model that used only cell line and drug identifiers. Molecular fingerprint-based drug representations performed slightly better than learned representations-ECFP4 fingerprints increased R2 by 5.3% and Spearman correlation by 2.8% w.r.t the best learned representations. In general, fully connected feature-encoding subnetworks outperformed other architectures. DL outperformed other ML methods by more than 35% (R2) and 14% (Spearman). Additionally, an ensemble combining the top DL and ML models improved performance by about 6.5% (R2) and 4% (Spearman). Using a state-of-the-art interpretability method, we showed that DL models can learn to associate drug and cell line features with drug response in a biologically meaningful way. The strategies explored in this study will help to improve the development of computational methods for the rational design of effective drug combinations for cancer therapy.

The first multi-tissue genome-scale metabolic model of a woody plant highlights suberin biosynthesis pathways in Quercus suber.

Over the last decade, genome-scale metabolic models have been increasingly used to study plant metabolic behaviour at the tissue and multi-tissue level under different environmental conditions. Quercus suber, also known as the cork oak tree, is one of the most important forest communities of the Mediterranean/Iberian region. In this work, we present the genome-scale metabolic model of the Q. suber (iEC7871). The metabolic model comprises 7871 genes, 6231 reactions, and 6481 metabolites across eight compartments. Transcriptomics data was integrated into the model to obtain tissue-specific models for the leaf, inner bark, and phellogen, with specific biomass compositions. The tissue-specific models were merged into a diel multi-tissue metabolic model to predict interactions among the three tissues at the light and dark phases. The metabolic models were also used to analyse the pathways associated with the synthesis of suberin monomers, namely the acyl-lipids, phenylpropanoids, isoprenoids, and flavonoids production. The models developed in this work provide a systematic overview of the metabolism of Q. suber, including its secondary metabolism pathways and cork formation.

merlin, an improved framework for the reconstruction of high-quality genome-scale metabolic models.

Genome-scale metabolic models have been recognised as useful tools for better understanding living organisms' metabolism. merlin (https://www.merlin-sysbio.org/) is an open-source and user-friendly resource that hastens the models' reconstruction process, conjugating manual and automatic procedures, while leveraging the user's expertise with a curation-oriented graphical interface. An updated and redesigned version of merlin is herein presented. Since 2015, several features have been implemented in merlin, along with deep changes in the software architecture, operational flow, and graphical interface. The current version (4.0) includes the implementation of novel algorithms and third-party tools for genome functional annotation, draft assembly, model refinement, and curation. Such updates increased the user base, resulting in multiple published works, including genome metabolic (re-)annotations and model reconstructions of multiple (lower and higher) eukaryotes and prokaryotes. merlin version 4.0 is the only tool able to perform template based and de novo draft reconstructions, while achieving competitive performance compared to state-of-the art tools both for well and less-studied organisms.

The Characterization and Cytotoxic Evaluation of Chondrosia reniformis Collagen Isolated from Different Body Parts (Ectosome and Choanosome) Envisaging the Development of Biomaterials.

Chondrosia reniformis is a collagen-rich marine sponge that is considered a sustainable and viable option for producing an alternative to mammalian-origin collagens. However, there is a lack of knowledge regarding the properties of collagen isolated from different sponge parts, namely the outer region, or cortex, (ectosome) and the inner region (choanosome), and how it affects the development of biomaterials. In this study, a brief histological analysis focusing on C. reniformis collagen spatial distribution and a comprehensive comparative analysis between collagen isolated from ectosome and choanosome are presented. The isolated collagen characterization was based on isolation yield, Fourier-transformed infrared spectroscopy (FTIR), circular dichroism (CD), SDS-PAGE, dot blot, and amino acid composition, as well as their cytocompatibility envisaging the development of future biomedical applications. An isolation yield of approximately 20% was similar for both sponge parts, as well as the FTIR, CD, and SDS-PAGE profiles, which demonstrated that both isolated collagens presented a high purity degree and preserved their triple helix and fibrillar conformation. Ectosome collagen had a higher OHpro content and possessed collagen type I and IV, while the choanosome was predominately constituted by collagen type IV. In vitro cytotoxicity assays using the L929 fibroblast cell line displayed a significant cytotoxic effect of choanosome collagen at 2 mg/mL, while ectosome collagen enhanced cell metabolism and proliferation, thus indicating the latter as being more suitable for the development of biomaterials. This research represents a unique comparative study of C. reniformis body parts, serving as a support for further establishing this marine sponge as a promising alternative collagen source for the future development of biomedical applications.

Deep learning for drug response prediction in cancer.

Predicting the sensitivity of tumors to specific anti-cancer treatments is a challenge of paramount importance for precision medicine. Machine learning(ML) algorithms can be trained on high-throughput screening data to develop models that are able to predict the response of cancer cell lines and patients to novel drugs or drug combinations. Deep learning (DL) refers to a distinct class of ML algorithms that have achieved top-level performance in a variety of fields, including drug discovery. These types of models have unique characteristics that may make them more suitable for the complex task of modeling drug response based on both biological and chemical data, but the application of DL to drug response prediction has been unexplored until very recently. The few studies that have been published have shown promising results, and the use of DL for drug response prediction is beginning to attract greater interest from researchers in the field. In this article, we critically review recently published studies that have employed DL methods to predict drug response in cancer cell lines. We also provide a brief description of DL and the main types of architectures that have been used in these studies. Additionally, we present a selection of publicly available drug screening data resources that can be used to develop drug response prediction models. Finally, we also address the limitations of these approaches and provide a discussion on possible paths for further improvement. Contact:  mrocha@di.uminho.pt.

A pipeline for the reconstruction and evaluation of context-specific human metabolic models at a large-scale.

Constraint-based (CB) metabolic models provide a mathematical framework and scaffold for in silico cell metabolism analysis and manipulation. In the past decade, significant efforts have been done to model human metabolism, enabled by the increased availability of multi-omics datasets and curated genome-scale reconstructions, as well as the development of several algorithms for context-specific model (CSM) reconstruction. Although CSM reconstruction has revealed insights on the deregulated metabolism of several pathologies, the process of reconstructing representative models of human tissues still lacks benchmarks and appropriate integrated software frameworks, since many tools required for this process are still disperse across various software platforms, some of which are proprietary. In this work, we address this challenge by assembling a scalable CSM reconstruction pipeline capable of integrating transcriptomics data in CB models. We combined omics preprocessing methods inspired by previous efforts with in-house implementations of existing CSM algorithms and new model refinement and validation routines, all implemented in the Troppo Python-based open-source framework. The pipeline was validated with multi-omics datasets from the Cancer Cell Line Encyclopedia (CCLE), also including reference fluxomics measurements for the MCF7 cell line. We reconstructed over 6000 models based on the Human-GEM template model for 733 cell lines featured in the CCLE, using MCF7 models as reference to find the best parameter combinations. These reference models outperform earlier studies using the same template by comparing gene essentiality and fluxomics experiments. We also analysed the heterogeneity of breast cancer cell lines, identifying key changes in metabolism related to cancer aggressiveness. Despite the many challenges in CB modelling, we demonstrate using our pipeline that combining transcriptomics data in metabolic models can be used to investigate key metabolic shifts. Significant limitations were found on these models ability for reliable quantitative flux prediction, thus motivating further work in genome-wide phenotype prediction.

Transitions Between Preexposure Prophylaxis Eligibility States and HIV Infection in the Lisbon Cohort of HIV-Negative Men Who Have Sex With Men: A Multistate Model Analysis.

We aimed to describe transitions between preexposure prophylaxis (PrEP) eligibility and human immunodeficiency virus (HIV) infection among HIV-negative men who have sex with men (MSM). We used data from 1,885 MSM, who had not used PrEP, enrolled in the Lisbon Cohort of MSM, with at least 2 consecutive measurements of PrEP eligibility from 2014-2020. A time-homogeneous Markov multistate model was applied to describe the transitions between states of PrEP eligibility-eligible and ineligible-and from these to HIV infection (HIV). The intensities of the transitions were closer for ineligible-to-eligible and eligible-to-ineligible transitions (intensity ratio, 1.107, 95% confidence interval (CI): 1.080, 1.176), while the intensity of the eligible-to-HIV transition was higher than that for ineligible-to-HIV transition (intensity ratio, 9.558, 95% CI: 0.738, 65.048). The probabilities of transitions increased with time; for 90 days, the probabilities were similar for the ineligible-to-eligible and eligible-to-ineligible transitions (0.285 (95% CI: 0.252, 0.319) vs. 0.258 (95% CI: 0.228, 0.287)), while the eligible-to-HIV transition was more likely than ineligible-to-HIV (0.004 (95% CI: 0.003, 0.007) vs. 0.001 (95% CI: 0.001, 0.008)) but tended to become closer with time. Being classified as ineligible was a short-term indicator of a lower probability of acquiring HIV. Once an individual moved to eligible, he was at a higher risk of seroconversion, demanding a timely delivery ofPrEP.

MEWpy: a computational strain optimization workbench in Python.

SUMMARY: Metabolic Engineering aims to favour the overproduction of native, as well as non-native, metabolites by modifying or extending the cellular processes of a specific organism. In this context, Computational Strain Optimization (CSO) plays a relevant role by putting forward mathematical approaches able to identify potential metabolic modifications to achieve the defined production goals. We present MEWpy, a Python workbench for metabolic engineering, which covers a wide range of metabolic and regulatory modelling approaches, as well as phenotype simulation and CSO algorithms. AVAILABILITY AND IMPLEMENTATION: MEWpy can be installed from PyPi (pip install mewpy), the source code being available at https://github.com/BioSystemsUM/mewpy under the GPL license.

Model-guided development of an evolutionarily stable yeast chassis.

First-principle metabolic modelling holds potential for designing microbial chassis that are resilient against phenotype reversal due to adaptive mutations. Yet, the theory of model-based chassis design has rarely been put to rigorous experimental test. Here, we report the development of Saccharomyces cerevisiae chassis strains for dicarboxylic acid production using genome-scale metabolic modelling. The chassis strains, albeit geared for higher flux towards succinate, fumarate and malate, do not appreciably secrete these metabolites. As predicted by the model, introducing product-specific TCA cycle disruptions resulted in the secretion of the corresponding acid. Adaptive laboratory evolution further improved production of succinate and fumarate, demonstrating the evolutionary robustness of the engineered cells. In the case of malate, multi-omics analysis revealed a flux bypass at peroxisomal malate dehydrogenase that was missing in the yeast metabolic model. In all three cases, flux balance analysis integrating transcriptomics, proteomics and metabolomics data confirmed the flux re-routing predicted by the model. Taken together, our modelling and experimental results have implications for the computer-aided design of microbial cell factories.

Causal integration of multi-omics data with prior knowledge to generate mechanistic hypotheses.

Multi-omics datasets can provide molecular insights beyond the sum of individual omics. Various tools have been recently developed to integrate such datasets, but there are limited strategies to systematically extract mechanistic hypotheses from them. Here, we present COSMOS (Causal Oriented Search of Multi-Omics Space), a method that integrates phosphoproteomics, transcriptomics, and metabolomics datasets. COSMOS combines extensive prior knowledge of signaling, metabolic, and gene regulatory networks with computational methods to estimate activities of transcription factors and kinases as well as network-level causal reasoning. COSMOS provides mechanistic hypotheses for experimental observations across multi-omics datasets. We applied COSMOS to a dataset comprising transcriptomics, phosphoproteomics, and metabolomics data from healthy and cancerous tissue from eleven clear cell renal cell carcinoma (ccRCC) patients. COSMOS was able to capture relevant crosstalks within and between multiple omics layers, such as known ccRCC drug targets. We expect that our freely available method will be broadly useful to extract mechanistic insights from multi-omics studies.

Biomaterials and Bioactive Natural Products from Marine Invertebrates: From Basic Research to Innovative Applications.

Aquatic invertebrates are a major source of biomaterials and bioactive natural products that can find applications as pharmaceutics, nutraceutics, cosmetics, antibiotics, antifouling products and biomaterials. Symbiotic microorganisms are often the real producers of many secondary metabolites initially isolated from marine invertebrates; however, a certain number of them are actually synthesized by the macro-organisms. In this review, we analysed the literature of the years 2010-2019 on natural products (bioactive molecules and biomaterials) from the main phyla of marine invertebrates explored so far, including sponges, cnidarians, molluscs, echinoderms and ascidians, and present relevant examples of natural products of interest to public and private stakeholders. We also describe omics tools that have been more relevant in identifying and understanding mechanisms and processes underlying the biosynthesis of secondary metabolites in marine invertebrates. Since there is increasing attention on finding new solutions for a sustainable large-scale supply of bioactive compounds, we propose that a possible improvement in the biodiscovery pipeline might also come from the study and utilization of aquatic invertebrate stem cells.

Going virtual and going wide: comparing Team-Based Learning in-class versus online and across disciplines.

Team-based learning (TBL) is an active learning pedagogy developed for in-class sessions and based on the collaborative work of small groups of students. The increasing push to online and blended learning has enhanced the need to expand this pedagogy to a virtual environment, but little evidence has been produced on how students accept online synchronous sessions of TBL. The purpose of this study, that relies on 427 responses, is to present a comparative perspective of traditional in-class versus adapted fully synchronous online TBL and across different disciplinary fields. Students of two different academic years and different programs were surveyed for their acceptance of TBL. They were invited to answer closed-ended questions focused on their engagement in all TBL learning process and the final outcomes provided. Results obtained from this unique comparative study revealed a wide approval of TBL, regardless of the environment (online or in-class TBL sessions), scientific area of courses and student gender. The acceptance of fully online TBL sessions, in a similar way as traditional in-class sessions, could be a rationale for giving more use to the 'virtual' context. Other results corroborated previous researches on TBL, such the need of student awareness of TBL benefits to get more engaged in the process or the impact of student activities overload on the TBL process. Implications are informative for pedagogical practice.

Transcontinental Dissemination of the L2b/D-Da Recombinant Chlamydia trachomatis Lymphogranuloma venereum (LGV) Strain: Need of Broad Multi-Country Molecular Surveillance.

Previously, we identified a Chlamydia trachomatis lymphogranuloma venereum (LGV) recombinant strain possessing a non-LGV ompA genotype. Here, culture-independent genome sequencing confirms its circulation in Europe, Middle East, and North America, and unveils emergence of antibiotic resistance. Broad surveillance is needed.

NMRFinder: a novel method for 1D 1H-NMR metabolite annotation.

INTRODUCTION: Methods for the automated and accurate identification of metabolites in 1D 1H-NMR samples are crucial, but this is still an unsolved problem. Most available tools are mainly focused on metabolite quantification, thus limiting the number of metabolites that can be identified. Also, most only use reference spectra obtained under the same specific conditions of the target sample, limiting the use of available knowledge. OBJECTIVES: The main goal of this work was to develop novel methods to perform metabolite annotation from 1D 1H-NMR peaks with enhanced reliability, to aid the users in metabolite identification. An essential step was to construct a vast and up-do-date library of reference 1D 1H-NMR peak lists collected under distinct experimental conditions. METHODS: Three different algorithms were evaluated for their capacity to correctly annotate metabolites present in both synthetic and real samples and compared to publicly available tools. The best proposed method was evaluated in a plethora of scenarios, including missing references, missing peaks and peak shifts, to assess its annotation accuracy, precision and recall. RESULTS: We gathered 1816 peak lists for 1387 different metabolites from several sources across different conditions for our reference library. A new method, NMRFinder, is proposed and allows matching 1D 1H-NMR samples with all the reference peak lists in the library, regardless of acquisition conditions. Metabolites are scored according to the number of peaks matching the samples, how unique their peaks are in the library and how close the spectrum acquisition conditions are in relation to those of the samples. Results show a true positive rate of 0.984 when analysing computationally created samples, while 71.8% of the metabolites were annotated when analysing samples from previously identified public datasets. CONCLUSION: NMRFinder performs metabolite annotation reliably and outperforms previous methods, being of great value in helping the user to ultimately identify metabolites. It is implemented in the R package specmine.

Diverse and Productive Source of Biopolymer Inspiration: Marine Collagens.

Marine biodiversity is expressed through the huge variety of vertebrate and invertebrate species inhabiting intertidal to deep-sea environments. The extraordinary variety of "forms and functions" exhibited by marine animals suggests they are a promising source of bioactive molecules and provides potential inspiration for different biomimetic approaches. This diversity is familiar to biologists and has led to intensive investigation of metabolites, polysaccharides, and other compounds. However, marine collagens are less well-known. This review will provide detailed insight into the diversity of collagens present in marine species in terms of their genetics, structure, properties, and physiology. In the last part of the review the focus will be on the most common marine collagen sources and on the latest advances in the development of innovative materials exploiting, or inspired by, marine collagens.

Provision of Preexposure Prophylaxis at the Portuguese National Health Service and Uptake in the Lisbon Cohort of Men Who Have Sex with Men.

We aimed to study the uptake of preexposure prophylaxis (PrEP) before and after its implementation in the Portuguese National Health Service (PNHS) among men who have sex with men (MSM). We studied 6164 participants in the Lisbon Cohort of MSM who participated between March 2014 and July 2019. 198 participants (3.2%) reported having recently used PrEP. Approximately one-third started PrEP after its implementation. PrEP uptake increased from 0.15% in 2014 to 5.36% in 2019. In their first use, 86 participants (70.5%) used it daily. How PrEP was obtained varied according to the timing of the first use: prescribed by a physician in Portugal (11.1% before vs 68.8% after implementation) and online (40.7% before vs 14.1% after). We observed an increase in the uptake and in the prescription by a physician, particularly after its implementation in the PNHS representing a change to a more equitable and safer way of using PrEP.

RESUMEN: Nuestro objetivo fue estudiar el uso de profilaxis pre exposición (PrEP) en hombres que tienen sexo con hombres (HSH), antes y después de su implementación en el Servicio Nacional de Salud de Portugal (PNHS). Estudiamos 6164 participantes de la Cohorte de HSH de Lisboa evaluados entre marzo de 2014 y julio de 2019. De los participantes, 198 (3,2%) indicaron haber utilizado PrEP recientemente. Aproximadamente un tercio comenzó con PrEP después de su implementación en el PNHS. El uso de PrEP aumentó del 0,15% en 2014 al 5,36% en 2019. En su primer uso, 86 participantes (70,5%) la usaron a diario. La forma en que se obtuvo la PrEP varió según el momento del primer uso: prescrito por un médico en Portugal (11,1% antes vs. 68,8% después de la implementación) y online (40,7% antes vs. 14,1% después). Observamos un aumento en el uso y en prescripción médica, particularmente después de su implementación en el PNHS, lo que representa un cambio hacia una forma más equitativa y segura de usar la PrEP.

Generative Deep Learning for Targeted Compound Design.

In the past few years, de novo molecular design has increasingly been using generative models from the emergent field of Deep Learning, proposing novel compounds that are likely to possess desired properties or activities. De novo molecular design finds applications in different fields ranging from drug discovery and materials sciences to biotechnology. A panoply of deep generative models, including architectures as Recurrent Neural Networks, Autoencoders, and Generative Adversarial Networks, can be trained on existing data sets and provide for the generation of novel compounds. Typically, the new compounds follow the same underlying statistical distributions of properties exhibited on the training data set Additionally, different optimization strategies, including transfer learning, Bayesian optimization, reinforcement learning, and conditional generation, can direct the generation process toward desired aims, regarding their biological activities, synthesis processes or chemical features. Given the recent emergence of these technologies and their relevance, this work presents a systematic and critical review on deep generative models and related optimization methods for targeted compound design, and their applications.

Use of Hanseniaspora guilliermondii and Hanseniaspora opuntiae to enhance the aromatic profile of beer in mixed-culture fermentation with Saccharomyces cerevisiae.

Beer production is predominantly carried out by Saccharomyces species, such as S. cerevisiae and S. pastorianus. However, the introduction of non-Saccharomyces yeasts in the brewing process is now seen as a promising strategy to improve and differentiate the organoleptic profile of beer. In this study, 17 non-Saccharomyces strains of 12 distinct species were isolated and submitted to a preliminary sensory evaluation to determine their potential for beer bioflavouring. Hanseniaspora guilliermondii IST315 and H. opuntiae IST408 aroma profiles presented the highest acceptability and were described as having 'fruity' and 'toffee' notes, respectively. Their presence in mixed-culture fermentations with S. cerevisiae US-05 did not influence attenuation and ethanol concentration of beer but had a significant impact in its volatile composition. Notably, while both strains reduced the total amount of ethyl esters, H. guilliermondii IST315 greatly increased the concentration of acetate esters, especially when sequentially inoculated, leading to an 8.2-fold increase in phenylethyl acetate ('rose', 'honey' aroma) in the final beverage. These findings highlight the importance of non-Saccharomyces yeasts in shaping the aroma profile of beer and suggest a role for Hanseniaspora spp. in improving it.

CoBAMP: a Python framework for metabolic pathway analysis in constraint-based models.

SUMMARY: CoBAMP is a modular framework for the enumeration of pathway analysis concepts, such as elementary flux modes (EFM) and minimal cut sets in genome-scale constraint-based models (CBMs) of metabolism. It currently includes the K-shortest EFM algorithm and facilitates integration with other frameworks involving reading, manipulation and analysis of CBMs. AVAILABILITY AND IMPLEMENTATION: The software is implemented in Python 3, supported on most operating systems and requires a mixed-integer linear programming optimizer supported by the optlang framework. Source-code is available at https://github.com/BioSystemsUM/cobamp.

Predicting promoters in phage genomes using PhagePromoter.

SUMMARY: The growing interest in phages as antibacterial agents has led to an increase in the number of sequenced phage genomes, increasing the need for intuitive bioinformatics tools for performing genome annotation. The identification of phage promoters is indeed the most difficult step of this process. Due to the lack of online tools for phage promoter prediction, we developed PhagePromoter, a tool for locating promoters in phage genomes, using machine learning methods. This is the first online tool for predicting promoters that uses phage promoter data and the first to identify both host and phage promoters with different motifs. AVAILABILITY AND IMPLEMENTATION: This tool was integrated in the Galaxy framework and it is available online at: https://bit.ly/2Dfebfv. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.