FSM-DDTR: End-to-end feedback strategy for multi-objective De Novo drug design using transformers.

The design of compounds that target specific biological functions with relevant selectivity is critical in the context of drug discovery, especially due to the polypharmacological nature of most existing drug molecules. In recent years, in silico-based methods combined with deep learning have shown promising results in the de novo drug design challenge, leading to potential leads for biologically interesting targets. However, several of these methods overlook the importance of certain properties, such as validity rate and target selectivity, or simplify the generative process by neglecting the multi-objective nature of the pharmacological space. In this study, we propose a multi-objective Transformer-based architecture to generate drug candidates with desired molecular properties and increased selectivity toward a specific biological target. The framework consists of a Transformer-Decoder Generator that generates novel and valid compounds in the SMILES format notation, a Transformer-Encoder Predictor that estimates the binding affinity toward the biological target, and a feedback loop combined with a multi-objective optimization strategy to rank the generated molecules and condition the generating distribution around the targeted properties. The results demonstrate that the proposed architecture can generate novel and synthesizable small compounds with desired pharmacological properties toward a biologically relevant target. The unbiased Transformer-based Generator achieved superior performance in the novelty rate (97.38%) and comparable performance in terms of internal diversity, uniqueness, and validity against state-of-the-art baselines. The optimization of the unbiased Transformer-based Generator resulted in the generation of molecules exhibiting high binding affinity toward the Adenosine A2A Receptor (AA2AR) and possessing desirable physicochemical properties, where 99.36% of the generated molecules follow Lipinski's rule of five. Furthermore, the implementation of a feedback strategy, in conjunction with a multi-objective algorithm, effectively shifted the distribution of the generated molecules toward optimal values of molecular weight, molecular lipophilicity, topological polar surface area, synthetic accessibility score, and quantitative estimate of drug-likeness, without the necessity of prior training sets comprising molecules endowed with pharmacological properties of interest. Overall, this research study validates the applicability of a Transformer-based architecture in the context of drug design, capable of exploring the vast chemical representation space to generate novel molecules with improved pharmacological properties and target selectivity. The data and source code used in this study are available at: https://github.com/larngroup/FSM-DDTR.

A Reliable and Secure Method for Sharing Genomic Data.

Genomics has significantly impacted the field of medicine, with advances in DNA sequencing leading to personalized medicine and a deeper understanding of the genomic basis of various diseases. The ability to share genomic data is crucial for advancing this field and developing new approaches to understanding the genome. However, the sensitive nature of this data requires secure methods for protecting it during storage and transfer. In this paper, we present a new tool for the secure encryption and decryption of FASTA files without sharing a common secret and with a reduced number of shared keys between the pairs. Our proposal combines symmetric and asymmetric encryption techniques, including the AES (Advanced Encryption Standard) cypher and RSA (Rivest-Shamir-Adleman). The tool is fast, reliable, and secure, outperforming existing tools in terms of security and ease of use. This makes it a valuable solution for the secure sharing and use of sensitive genomic data, representing a significant advancement in the field of genomics.

DTITR: End-to-end drug-target binding affinity prediction with transformers.

The accurate identification of Drug-Target Interactions (DTIs) remains a critical turning point in drug discovery and understanding of the binding process. Despite recent advances in computational solutions to overcome the challenges of in vitro and in vivo experiments, most of the proposed in silico-based methods still focus on binary classification, overlooking the importance of characterizing DTIs with unbiased binding strength values to properly distinguish primary interactions from those with off-targets. Moreover, several of these methods usually simplify the entire interaction mechanism, neglecting the joint contribution of the individual units of each binding component and the interacting substructures involved, and have yet to focus on more explainable and interpretable architectures. In this study, we propose an end-to-end Transformer-based architecture for predicting drug-target binding affinity (DTA) using 1D raw sequential and structural data to represent the proteins and compounds. This architecture exploits self-attention layers to capture the biological and chemical context of the proteins and compounds, respectively, and cross-attention layers to exchange information and capture the pharmacological context of the DTIs. The results show that the proposed architecture is effective in predicting DTA, achieving superior performance in both correctly predicting the value of interaction strength and being able to correctly discriminate the rank order of binding strength compared to state-of-the-art baselines. The combination of multiple Transformer-Encoders was found to result in robust and discriminative aggregate representations of the proteins and compounds for binding affinity prediction, in which the addition of a Cross-Attention Transformer-Encoder was identified as an important block for improving the discriminative power of these representations. Overall, this research study validates the applicability of an end-to-end Transformer-based architecture in the context of drug discovery, capable of self-providing different levels of potential DTI and prediction understanding due to the nature of the attention blocks. The data and source code used in this study are available at: https://github.com/larngroup/DTITR.

Explainable deep drug-target representations for binding affinity prediction.

BACKGROUND: Several computational advances have been achieved in the drug discovery field, promoting the identification of novel drug-target interactions and new leads. However, most of these methodologies have been overlooking the importance of providing explanations to the decision-making process of deep learning architectures. In this research study, we explore the reliability of convolutional neural networks (CNNs) at identifying relevant regions for binding, specifically binding sites and motifs, and the significance of the deep representations extracted by providing explanations to the model's decisions based on the identification of the input regions that contributed the most to the prediction. We make use of an end-to-end deep learning architecture to predict binding affinity, where CNNs are exploited in their capacity to automatically identify and extract discriminating deep representations from 1D sequential and structural data. RESULTS: The results demonstrate the effectiveness of the deep representations extracted from CNNs in the prediction of drug-target interactions. CNNs were found to identify and extract features from regions relevant for the interaction, where the weight associated with these spots was in the range of those with the highest positive influence given by the CNNs in the prediction. The end-to-end deep learning model achieved the highest performance both in the prediction of the binding affinity and on the ability to correctly distinguish the interaction strength rank order when compared to baseline approaches. CONCLUSIONS: This research study validates the potential applicability of an end-to-end deep learning architecture in the context of drug discovery beyond the confined space of proteins and ligands with determined 3D structure. Furthermore, it shows the reliability of the deep representations extracted from the CNNs by providing explainability to the decision-making process.

Designing optimized drug candidates with Generative Adversarial Network.

Drug design is an important area of study for pharmaceutical businesses. However, low efficacy, off-target delivery, time consumption, and high cost are challenges and can create barriers that impact this process. Deep Learning models are emerging as a promising solution to perform de novo drug design, i.e., to generate drug-like molecules tailored to specific needs. However, stereochemistry was not explicitly considered in the generated molecules, which is inevitable in targeted-oriented molecules. This paper proposes a framework based on Feedback Generative Adversarial Network (GAN) that includes optimization strategy by incorporating Encoder-Decoder, GAN, and Predictor deep models interconnected with a feedback loop. The Encoder-Decoder converts the string notations of molecules into latent space vectors, effectively creating a new type of molecular representation. At the same time, the GAN can learn and replicate the training data distribution and, therefore, generate new compounds. The feedback loop is designed to incorporate and evaluate the generated molecules according to the multiobjective desired property at every epoch of training to ensure a steady shift of the generated distribution towards the space of the targeted properties. Moreover, to develop a more precise set of molecules, we also incorporate a multiobjective optimization selection technique based on a non-dominated sorting genetic algorithm. The results demonstrate that the proposed framework can generate realistic, novel molecules that span the chemical space. The proposed Encoder-Decoder model correctly reconstructs 99% of the datasets, including stereochemical information. The model's ability to find uncharted regions of the chemical space was successfully shown by optimizing the unbiased GAN to generate molecules with a high binding affinity to the Kappa Opioid and Adenosine [Formula: see text] receptor. Furthermore, the generated compounds exhibit high internal and external diversity levels 0.88 and 0.94, respectively, and uniqueness.

Handling Noise in Protein Interaction Networks.

Protein-protein interactions (PPIs) can be conveniently represented as networks, allowing the use of graph theory for their study. Network topology studies may reveal patterns associated with specific organisms. Here, we propose a new methodology to denoise PPI networks and predict missing links solely based on the network topology, the organization measurement (OM) method. The OM methodology was applied in the denoising of the PPI networks of two Saccharomyces cerevisiae datasets (Yeast and CS2007) and one Homo sapiens dataset (Human). To evaluate the denoising capabilities of the OM methodology, two strategies were applied. The first strategy compared its application in random networks and in the reference set networks, while the second strategy perturbed the networks with the gradual random addition and removal of edges. The application of the OM methodology to the Yeast and Human reference sets achieved an AUC of 0.95 and 0.87, in Yeast and Human networks, respectively. The random removal of 80% of the Yeast and Human reference set interactions resulted in an AUC of 0.71 and 0.62, whereas the random addition of 80% interactions resulted in an AUC of 0.75 and 0.72, respectively. Applying the OM methodology to the CS2007 dataset yields an AUC of 0.99. We also perturbed the network of the CS2007 dataset by randomly inserting and removing edges in the same proportions previously described. The false positives identified and removed from the network varied from 97%, when inserting 20% more edges, to 89%, when 80% more edges were inserted. The true positives identified and inserted in the network varied from 95%, when removing 20% of the edges, to 40%, after the random deletion of 80% edges. The OM methodology is sensitive to the topological structure of the biological networks. The obtained results suggest that the present approach can efficiently be used to denoise PPI networks.

Correction: Phenotypic heterogeneity promotes adaptive evolution.

[This corrects the article DOI: 10.1371/journal.pbio.2000644.].

Phenotypic heterogeneity promotes adaptive evolution.

Genetically identical cells frequently display substantial heterogeneity in gene expression, cellular morphology and physiology. It has been suggested that by rapidly generating a subpopulation with novel phenotypic traits, phenotypic heterogeneity (or plasticity) accelerates the rate of adaptive evolution in populations facing extreme environmental challenges. This issue is important as cell-to-cell phenotypic heterogeneity may initiate key steps in microbial evolution of drug resistance and cancer progression. Here, we study how stochastic transitions between cellular states influence evolutionary adaptation to a stressful environment in yeast Saccharomyces cerevisiae. We developed inducible synthetic gene circuits that generate varying degrees of expression stochasticity of an antifungal resistance gene. We initiated laboratory evolutionary experiments with genotypes carrying different versions of the genetic circuit by exposing the corresponding populations to gradually increasing antifungal stress. Phenotypic heterogeneity altered the evolutionary dynamics by transforming the adaptive landscape that relates genotype to fitness. Specifically, it enhanced the adaptive value of beneficial mutations through synergism between cell-to-cell variability and genetic variation. Our work demonstrates that phenotypic heterogeneity is an evolving trait when populations face a chronic selection pressure. It shapes evolutionary trajectories at the genomic level and facilitates evolutionary rescue from a deteriorating environmental stress.

A Cloud Architecture for Teleradiology-as-a-Service.

BACKGROUND: Telemedicine has been promoted by healthcare professionals as an efficient way to obtain remote assistance from specialised centres, to get a second opinion about complex diagnosis or even to share knowledge among practitioners. The current economic restrictions in many countries are increasing the demand for these solutions even more, in order to optimize processes and reduce costs. However, despite some technological solutions already in place, their adoption has been hindered by the lack of usability, especially in the set-up process. OBJECTIVES: In this article we propose a telemedicine platform that relies on a cloud computing infrastructure and social media principles to simplify the creation of dynamic user-based groups, opening up opportunities for the establishment of teleradiology trust domains. METHODS: The collaborative platform is provided as a Software-as-a-Service solution, supporting real time and asynchronous collaboration between users. To evaluate the solution, we have deployed the platform in a private cloud infrastructure. The system is made up of three main components - the collaborative framework, the Medical Management Information System (MMIS) and the HTML5 (Hyper Text Markup Language) Web client application - connected by a message-oriented middleware. RESULTS: The solution allows physicians to create easily dynamic network groups for synchronous or asynchronous cooperation. The network created improves dataflow between colleagues and also knowledge sharing and cooperation through social media tools. The platform was implemented and it has already been used in two distinct scenarios: teaching of radiology and tele-reporting. CONCLUSIONS: Collaborative systems can simplify the establishment of telemedicine expert groups with tools that enable physicians to improve their clinical practice. Streamlining the usage of this kind of systems through the adoption of Web technologies that are common in social media will increase the quality of current solutions, facilitating the sharing of clinical information, medical imaging studies and patient diagnostics among collaborators.

Representation learning for mammography mass lesion classification with convolutional neural networks.

BACKGROUND AND OBJECTIVE: The automatic classification of breast imaging lesions is currently an unsolved problem. This paper describes an innovative representation learning framework for breast cancer diagnosis in mammography that integrates deep learning techniques to automatically learn discriminative features avoiding the design of specific hand-crafted image-based feature detectors. METHODS: A new biopsy proven benchmarking dataset was built from 344 breast cancer patients' cases containing a total of 736 film mammography (mediolateral oblique and craniocaudal) views, representative of manually segmented lesions associated with masses: 426 benign lesions and 310 malignant lesions. The developed method comprises two main stages: (i) preprocessing to enhance image details and (ii) supervised training for learning both the features and the breast imaging lesions classifier. In contrast to previous works, we adopt a hybrid approach where convolutional neural networks are used to learn the representation in a supervised way instead of designing particular descriptors to explain the content of mammography images. RESULTS: Experimental results using the developed benchmarking breast cancer dataset demonstrated that our method exhibits significant improved performance when compared to state-of-the-art image descriptors, such as histogram of oriented gradients (HOG) and histogram of the gradient divergence (HGD), increasing the performance from 0.787 to 0.822 in terms of the area under the ROC curve (AUC). Interestingly, this model also outperforms a set of hand-crafted features that take advantage of additional information from segmentation by the radiologist. Finally, the combination of both representations, learned and hand-crafted, resulted in the best descriptor for mass lesion classification, obtaining 0.826 in the AUC score. CONCLUSIONS: A novel deep learning based framework to automatically address classification of breast mass lesions in mammography was developed.

An Overview of Biomolecular Event Extraction from Scientific Documents.

This paper presents a review of state-of-the-art approaches to automatic extraction of biomolecular events from scientific texts. Events involving biomolecules such as genes, transcription factors, or enzymes, for example, have a central role in biological processes and functions and provide valuable information for describing physiological and pathogenesis mechanisms. Event extraction from biomedical literature has a broad range of applications, including support for information retrieval, knowledge summarization, and information extraction and discovery. However, automatic event extraction is a challenging task due to the ambiguity and diversity of natural language and higher-level linguistic phenomena, such as speculations and negations, which occur in biological texts and can lead to misunderstanding or incorrect interpretation. Many strategies have been proposed in the last decade, originating from different research areas such as natural language processing, machine learning, and statistics. This review summarizes the most representative approaches in biomolecular event extraction and presents an analysis of the current state of the art and of commonly used methods, features, and tools. Finally, current research trends and future perspectives are also discussed.

A document processing pipeline for annotating chemical entities in scientific documents.

BACKGROUND: The recognition of drugs and chemical entities in text is a very important task within the field of biomedical information extraction, given the rapid growth in the amount of published texts (scientific papers, patents, patient records) and the relevance of these and other related concepts. If done effectively, this could allow exploiting such textual resources to automatically extract or infer relevant information, such as drug profiles, relations and similarities between drugs, or associations between drugs and potential drug targets. The objective of this work was to develop and validate a document processing and information extraction pipeline for the identification of chemical entity mentions in text. RESULTS: We used the BioCreative IV CHEMDNER task data to train and evaluate a machine-learning based entity recognition system. Using a combination of two conditional random field models, a selected set of features, and a post-processing stage, we achieved F-measure results of 87.48% in the chemical entity mention recognition task and 87.75% in the chemical document indexing task. CONCLUSIONS: We present a machine learning-based solution for automatic recognition of chemical and drug names in scientific documents. The proposed approach applies a rich feature set, including linguistic, orthographic, morphological, dictionary matching and local context features. Post-processing modules are also integrated, performing parentheses correction, abbreviation resolution and filtering erroneous mentions using an exclusion list derived from the training data. The developed methods were implemented as a document annotation tool and web service, freely available at http://bioinformatics.ua.pt/becas-chemicals/.

Convolutional neural networks for mammography mass lesion classification.

Feature extraction is a fundamental step when mammography image analysis is addressed using learning based approaches. Traditionally, problem dependent handcrafted features are used to represent the content of images. An alternative approach successfully applied in other domains is the use of neural networks to automatically discover good features. This work presents an evaluation of convolutional neural networks to learn features for mammography mass lesions before feeding them to a classification stage. Experimental results showed that this approach is a suitable strategy outperforming the state-of-the-art representation from 79.9% to 86% in terms of area under the ROC curve.

Improving the accuracy of recombinant protein production through integration of bioinformatics, statistical and mass spectrometry methodologies.

Heterologous protein production is a key technology for biotechnological, health sciences and many other research fields. Various approaches have been developed for its optimization, but the research emphasis has been on optimization of protein yield rather than protein quality. In this study, we have established a workflow for synthetic gene optimization for heterologous protein expression that combines bioinformatics, laboratory experiments, mass spectrometry and statistical analysis. Two gene primary structure analysis platforms, Anaconda and EuGene, and multivariate optimization methods were employed to re-design the Plasmodium falciparum lysyl-tRNA synthetase gene for optimal expression in Escherichia coli. Synthetic genes were expressed from common vectors, and amino acid mis-incorporations in the expressed proteins were detected and quantified using mass spectrometry. The association between the identified amino acid mis-incorporations and 23 gene variables was then analysed. The synthetic genes yielded significantly higher levels of protein relative to the wild-type gene, but 71 amino acid mis-incorporation sites were observed along the whole protein and across the synthetic genes that were statistically associated with specific codons and protein secondary structures. The optimization method that led to production of the most accurate protein was based on a multivariate approach that combined variables that are known to influence mRNA translation.

Automatic filtering and substantiation of drug safety signals.

Drug safety issues pose serious health threats to the population and constitute a major cause of mortality worldwide. Due to the prominent implications to both public health and the pharmaceutical industry, it is of great importance to unravel the molecular mechanisms by which an adverse drug reaction can be potentially elicited. These mechanisms can be investigated by placing the pharmaco-epidemiologically detected adverse drug reaction in an information-rich context and by exploiting all currently available biomedical knowledge to substantiate it. We present a computational framework for the biological annotation of potential adverse drug reactions. First, the proposed framework investigates previous evidences on the drug-event association in the context of biomedical literature (signal filtering). Then, it seeks to provide a biological explanation (signal substantiation) by exploring mechanistic connections that might explain why a drug produces a specific adverse reaction. The mechanistic connections include the activity of the drug, related compounds and drug metabolites on protein targets, the association of protein targets to clinical events, and the annotation of proteins (both protein targets and proteins associated with clinical events) to biological pathways. Hence, the workflows for signal filtering and substantiation integrate modules for literature and database mining, in silico drug-target profiling, and analyses based on gene-disease networks and biological pathways. Application examples of these workflows carried out on selected cases of drug safety signals are discussed. The methodology and workflows presented offer a novel approach to explore the molecular mechanisms underlying adverse drug reactions.

Species-specific codon context rules unveil non-neutrality effects of synonymous mutations.

BACKGROUND: Codon pair usage (codon context) is a species specific gene primary structure feature whose evolutionary and functional roles are poorly understood. The data available show that codon-context has direct impact on both translation accuracy and efficiency, but one does not yet understand how it affects these two translation variables or whether context biases shape gene evolution. METHODOLOGIES/PRINCIPAL FINDINGS: Here we study codon-context biases using a set of 72 orthologous highly conserved genes from bacteria, archaea, fungi and high eukaryotes to identify 7 distinct groups of codon context rules. We show that synonymous mutations, i.e., neutral mutations that occur in synonymous codons of codon-pairs, are selected to maintain context biases and that non-synonymous mutations, i.e., non-neutral mutations that alter protein amino acid sequences, are also under selective pressure to preserve codon-context biases. CONCLUSIONS: Since in vivo studies provide evidence for a role of codon context on decoding fidelity in E. coli and for decoding efficiency in mammalian cells, our data support the hypothesis that, like codon usage, codon context modulates the evolution of gene primary structure and fine tunes the structure of open reading frames for high genome translational fidelity and efficiency in the 3 domains of life.

Can tricuspid annuloplasty of the donor heart reduce valve insufficiency following cardiac transplantation with bicaval anastomosis?

BACKGROUND: The aim of this study was to evaluate the degree of tricuspid valve insufficiency after orthotopic cardiac transplantation with bicaval anastomosis and prophylactic donor heart annuloplasty. METHODS: At present, our cardiac transplantation experience includes 478 cases. After January 2002, we included 30 consecutive patients in this study who had undergone orthotopic cardiac transplantation and survived >6 months. The patients were divided into 2 groups: group I, 15 patients who underwent transplantation with prophylactic tricuspid annuloplasty on the donor heart with the De Vega technique; and group II, 15 patients who underwent transplantation without this procedure. Their preoperative clinical characteristics were the same. During the late postoperative follow-up, the degree of tricuspid insufficiency was evaluated by transthoracic Doppler echocardiography and assessed according to the Simpson scale: 0, absent; 1, mild; 2, moderate; and 3, severe. Hemodynamic parameters were evaluated invasively by means of a Swan-Ganz catheter during routine endomyocardial biopsies. RESULTS: The mean follow-up time was 26.9 +/- 5.4 months (range, 12-36 months). In group I, 1 patient (6.6%) died from infection in the 18th month after the operation; the death was not related to the annuloplasty. In group II, 1 death (6.6%) occurred after 10 months because of rejection (P > .05). After the 24-month follow-up, the mean degree of tricuspid insufficiency was 0.4 +/- 0.5 in group I and 1.7 +/- 0.9 in group II (P < .05). Similarly, the 2 groups were significantly different with respect to the right atrium pressure, which was higher in group II. CONCLUSIONS: Prophylactic tricuspid annuloplasty on the donor heart was able to reduce significantly the degree of valvular insufficiency, even in cardiac transplantation with bicaval anastomosis; however, it did not modify significantly the hemodynamic performance of the allograft during the investigation period. It is very important to extend the observation period and casuistics to verify other benefits that this technique may offer.

Indexing and retrieving DICOM data in disperse and unstructured archives.

OBJECTIVE: This paper proposes an indexing and retrieval solution to gather information from distributed DICOM documents by allowing searches and access to the virtual data repository using a Google-like process. METHODS AND MATERIALS: The medical imaging modalities are becoming more powerful and less expensive. The result is the proliferation of equipment acquisition by imaging centers, including the small ones. With this dispersion of data, it is not easy to take advantage of all the information that can be retrieved from these studies. Furthermore, many of these small centers do not have large enough requirements to justify the acquisition of a traditional PACS. RESULTS: A peer-to-peer PACS platform to index and query DICOM files over a set of distributed repositories that are logically viewed as a single federated unit. The solution is based on a public domain document-indexing engine and extends traditional PACS query and retrieval mechanisms. CONCLUSION: This proposal deals well with complex searching requirements, from a single desktop environment to distributed scenarios. The solution performance and robustness were demonstrated in trials. The characteristics of presented PACS platform make it particularly important for small institutions, including educational and research groups.