A decision support system based on artificial intelligence and systems biology for the simulation of pancreatic cancer patient status.

Oncology treatments require continuous individual adjustment based on the measurement of multiple clinical parameters. Prediction tools exploiting the patterns present in the clinical data could be used to assist decision making and ease the burden associated to the interpretation of all these parameters. The goal of this study was to predict the evolution of patients with pancreatic cancer at their next visit using information routinely recorded in health records, providing a decision-support system for clinicians. We selected hematological variables as the visit's clinical outcomes, under the assumption that they can be predictive of the evolution of the patient. Multivariate models based on regression trees were generated to predict next-visit values for each of the clinical outcomes selected, based on the longitudinal clinical data as well as on molecular data sets streaming from in silico simulations of individual patient status at each visit. The models predict, with a mean prediction score (balanced accuracy) of 0.79, the evolution trends of eosinophils, leukocytes, monocytes, and platelets. Time span between visits and neutropenia were among the most common factors contributing to the predicted evolution. The inclusion of molecular variables from the systems-biology in silico simulations provided a molecular background for the observed variations in the selected outcome variables, mostly in relation to the regulation of hematopoiesis. In spite of its limitations, this study serves as a proof of concept for the application of next-visit prediction tools in real-world settings, even when available data sets are small.

Identification of biomarkers predictive of metastasis development in early-stage colorectal cancer using network-based regularization.

Colorectal cancer (CRC) is the third most common cancer and the second most deathly worldwide. It is a very heterogeneous disease that can develop via distinct pathways where metastasis is the primary cause of death. Therefore, it is crucial to understand the molecular mechanisms underlying metastasis. RNA-sequencing is an essential tool used for studying the transcriptional landscape. However, the high-dimensionality of gene expression data makes selecting novel metastatic biomarkers problematic. To distinguish early-stage CRC patients at risk of developing metastasis from those that are not, three types of binary classification approaches were used: (1) classification methods (decision trees, linear and radial kernel support vector machines, logistic regression, and random forest) using differentially expressed genes (DEGs) as input features; (2) regularized logistic regression based on the Elastic Net penalty and the proposed iTwiner-a network-based regularizer accounting for gene correlation information; and (3) classification methods based on the genes pre-selected using regularized logistic regression. Classifiers using the DEGs as features showed similar results, with random forest showing the highest accuracy. Using regularized logistic regression on the full dataset yielded no improvement in the methods' accuracy. Further classification using the pre-selected genes found by different penalty factors, instead of the DEGs, significantly improved the accuracy of the binary classifiers. Moreover, the use of network-based correlation information (iTwiner) for gene selection produced the best classification results and the identification of more stable and robust gene sets. Some are known to be tumor suppressor genes (OPCML-IT2), to be related to resistance to cancer therapies (RAC1P3), or to be involved in several cancer processes such as genome stability (XRCC6P2), tumor growth and metastasis (MIR602) and regulation of gene transcription (NME2P2). We show that the classification of CRC patients based on pre-selected features by regularized logistic regression is a valuable alternative to using DEGs, significantly increasing the models' predictive performance. Moreover, the use of correlation-based penalization for biomarker selection stands as a promising strategy for predicting patients' groups based on RNA-seq data.

Genetic and microenvironmental intra-tumor heterogeneity impacts colorectal cancer evolution and metastatic development.

Colorectal cancer (CRC) is a highly diverse disease, where different genomic instability pathways shape genetic clonal diversity and tumor microenvironment. Although intra-tumor heterogeneity has been characterized in primary tumors, its origin and consequences in CRC outcome is not fully understood. Therefore, we assessed intra- and inter-tumor heterogeneity of a prospective cohort of 136 CRC samples. We demonstrate that CRC diversity is forged by asynchronous forms of molecular alterations, where mutational and chromosomal instability collectively boost CRC genetic and microenvironment intra-tumor heterogeneity. We were able to depict predictor signatures of cancer-related genes that can foresee heterogeneity levels across the different tumor consensus molecular subtypes (CMS) and primary tumor location. Finally, we show that high genetic and microenvironment heterogeneity are associated with lower metastatic potential, whereas late-emerging copy number variations favor metastasis development and polyclonal seeding. This study provides an exhaustive portrait of the interplay between genetic and microenvironment intra-tumor heterogeneity across CMS subtypes, depicting molecular events with predictive value of CRC progression and metastasis development.

SYNPRED: prediction of drug combination effects in cancer using different synergy metrics and ensemble learning.

BACKGROUND: In cancer research, high-throughput screening technologies produce large amounts of multiomics data from different populations and cell types. However, analysis of such data encounters difficulties due to disease heterogeneity, further exacerbated by human biological complexity and genomic variability. The specific profile of cancer as a disease (or, more realistically, a set of diseases) urges the development of approaches that maximize the effect while minimizing the dosage of drugs. Now is the time to redefine the approach to drug discovery, bringing an artificial intelligence (AI)-powered informational view that integrates the relevant scientific fields and explores new territories. RESULTS: Here, we show SYNPRED, an interdisciplinary approach that leverages specifically designed ensembles of AI algorithms, as well as links omics and biophysical traits to predict anticancer drug synergy. It uses 5 reference models (Bliss, Highest Single Agent, Loewe, Zero Interaction Potency, and Combination Sensitivity Score), which, coupled with AI algorithms, allowed us to attain the ones with the best predictive performance and pinpoint the most appropriate reference model for synergy prediction, often overlooked in similar studies. By using an independent test set, SYNPRED exhibits state-of-the-art performance metrics either in the classification (accuracy, 0.85; precision, 0.91; recall, 0.90; area under the receiver operating characteristic, 0.80; and F1-score, 0.91) or in the regression models, mainly when using the Combination Sensitivity Score synergy reference model (root mean square error, 11.07; mean squared error, 122.61; Pearson, 0.86; mean absolute error, 7.43; Spearman, 0.87). Moreover, data interpretability was achieved by deploying the most current and robust feature importance approaches. A simple web-based application was constructed, allowing easy access by nonexpert researchers. CONCLUSIONS: The performance of SYNPRED rivals that of the existing methods that tackle the same problem, yielding unbiased results trained with one of the most comprehensive datasets available (NCI ALMANAC). The leveraging of different reference models allowed deeper insights into which of them can be more appropriately used for synergy prediction. The Combination Sensitivity Score clearly stood out with improved performance among the full scope of surveyed approaches and synergy reference models. Furthermore, SYNPRED takes a particular focus on data interpretability, which has been in the spotlight lately when using the most advanced AI techniques.

MENSAdb: a thorough structural analysis of membrane protein dimers.

Membrane proteins (MPs) are key players in a variety of different cellular processes and constitute the target of around 60% of all Food and Drug Administration-approved drugs. Despite their importance, there is still a massive lack of relevant structural, biochemical and mechanistic information mainly due to their localization within the lipid bilayer. To help fulfil this gap, we developed the MEmbrane protein dimer Novel Structure Analyser database (MENSAdb). This interactive web application summarizes the evolutionary and physicochemical properties of dimeric MPs to expand the available knowledge on the fundamental principles underlying their formation. Currently, MENSAdb contains features of 167 unique MPs (63% homo- and 37% heterodimers) and brings insights into the conservation of residues, accessible solvent area descriptors, average B-factors, intermolecular contacts at 2.5 Å and 4.0 Å distance cut-offs, hydrophobic contacts, hydrogen bonds, salt bridges, π-π stacking, T-stacking and cation-π interactions. The regular update and organization of all these data into a unique platform will allow a broad community of researchers to collect and analyse a large number of features efficiently, thus facilitating their use in the development of prediction models associated with MPs. Database URL: http://www.moreiralab.com/resources/mensadb.

HERVs establish a distinct molecular subtype in stage II/III colorectal cancer with poor outcome.

Colorectal cancer (CRC) is one of the most lethal malignancies. The extreme heterogeneity in survival rate is driving the need for new prognostic biomarkers. Human endogenous retroviruses (hERVs) have been suggested to influence tumor progression, oncogenesis and elicit an immune response. We examined multiple next-generation sequencing (NGS)-derived biomarkers in 114 CRC patients with paired whole-exome and whole-transcriptome sequencing (WES and WTS, respectively). First, we demonstrate that the median expression of hERVs can serve as a potential biomarker for prognosis, relapse, and resistance to chemotherapy in stage II and III CRC. We show that hERV expression and CD8+ tumor-infiltrating T-lymphocytes (TILs) synergistically stratify overall and relapse-free survival (OS and RFS): the median OS of the CD8-/hERV+ subgroup was 29.8 months compared with 37.5 months for other subgroups (HR = 4.4, log-rank P < 0.001). Combing NGS-based biomarkers (hERV/CD8 status) with clinicopathological factors provided a better prediction of patient survival compared to clinicopathological factors alone. Moreover, we explored the association between genomic and transcriptomic features of tumors with high hERV expression and establish this subtype as distinct from previously described consensus molecular subtypes of CRC. Overall, our results underscore a previously unknown role for hERVs in leading to a more aggressive subtype of CRC.

Predicting Hot Spots Using a Deep Neural Network Approach.

Targeting protein-protein interactions is a challenge and crucial task of the drug discovery process. A good starting point for rational drug design is the identification of hot spots (HS) at protein-protein interfaces, typically conserved residues that contribute most significantly to the binding. In this chapter, we depict point-by-point an in-house pipeline used for HS prediction using only sequence-based features from the well-known SpotOn dataset of soluble proteins (Moreira et al., Sci Rep 7:8007, 2017), through the implementation of a deep neural network. The presented pipeline is divided into three steps: (1) feature extraction, (2) deep learning classification, and (3) model evaluation. We present all the available resources, including code snippets, the main dataset, and the free and open-source modules/packages necessary for full replication of the protocol. The users should be able to develop an HS prediction model with accuracy, precision, recall, and AUROC of 0.96, 0.93, 0.91, and 0.86, respectively.

In-silico drug repurposing study predicts the combination of pirfenidone and melatonin as a promising candidate therapy to reduce SARS-CoV-2 infection progression and respiratory distress caused by cytokine storm.

From January 2020, COVID-19 is spreading around the world producing serious respiratory symptoms in infected patients that in some cases can be complicated by the severe acute respiratory syndrome, sepsis and septic shock, multiorgan failure, including acute kidney injury and cardiac injury. Cost and time efficient approaches to reduce the burthen of the disease are needed. To find potential COVID-19 treatments among the whole arsenal of existing drugs, we combined system biology and artificial intelligence-based approaches. The drug combination of pirfenidone and melatonin has been identified as a candidate treatment that may contribute to reduce the virus infection. Starting from different drug targets the effect of the drugs converges on human proteins with a known role in SARS-CoV-2 infection cycle. Simultaneously, GUILDify v2.0 web server has been used as an alternative method to corroborate the effect of pirfenidone and melatonin against the infection of SARS-CoV-2. We have also predicted a potential therapeutic effect of the drug combination over the respiratory associated pathology, thus tackling at the same time two important issues in COVID-19. These evidences, together with the fact that from a medical point of view both drugs are considered safe and can be combined with the current standard of care treatments for COVID-19 makes this combination very attractive for treating patients at stage II, non-severe symptomatic patients with the presence of virus and those patients who are at risk of developing severe pulmonary complications.

Prediction and targeting of GPCR oligomer interfaces.

GPCR oligomerization has emerged as a hot topic in the GPCR field in the last years. Receptors that are part of these oligomers can influence each other's function, although it is not yet entirely understood how these interactions work. The existence of such a highly complex network of interactions between GPCRs generates the possibility of alternative targets for new therapeutic approaches. However, challenges still exist in the characterization of these complexes, especially at the interface level. Different experimental approaches, such as FRET or BRET, are usually combined to study GPCR oligomer interactions. Computational methods have been applied as a useful tool for retrieving information from GPCR sequences and the few X-ray-resolved oligomeric structures that are accessible, as well as for predicting new and trustworthy GPCR oligomeric interfaces. Machine-learning (ML) approaches have recently helped with some hindrances of other methods. By joining and evaluating multiple structure-, sequence- and co-evolution-based features on the same algorithm, it is possible to dilute the issues of particular structures and residues that arise from the experimental methodology into all-encompassing algorithms capable of accurately predict GPCR-GPCR interfaces. All these methods used as a single or a combined approach provide useful information about GPCR oligomerization and its role in GPCR function and dynamics. Altogether, we present experimental, computational and machine-learning methods used to study oligomers interfaces, as well as strategies that have been used to target these dynamic complexes.

An Overview of Antiretroviral Agents for Treating HIV Infection in Paediatric Population.

Paediatric Acquired ImmunoDeficiency Syndrome (AIDS) is a life-threatening and infectious disease in which the Human Immunodeficiency Virus (HIV) is mainly transmitted through Mother-To- Child Transmission (MTCT) during pregnancy, labour and delivery, or breastfeeding. This review provides an overview of the distinct therapeutic alternatives to abolish the systemic viral replication in paediatric HIV-1 infection. Numerous classes of antiretroviral agents have emerged as therapeutic tools for downregulation of different steps in the HIV replication process. These classes encompass Non- Nucleoside Analogue Reverse Transcriptase Inhibitors (NNRTIs), Nucleoside/Nucleotide Analogue Reverse Transcriptase Inhibitors (NRTIs/NtRTIs), INtegrase Inhibitors (INIs), Protease Inhibitors (PIs), and Entry Inhibitors (EIs). Co-administration of certain antiretroviral drugs with Pharmacokinetic Enhancers (PEs) may boost the effectiveness of the primary therapeutic agent. The combination of multiple antiretroviral drug regimens (Highly Active AntiRetroviral Therapy - HAART) is currently the standard therapeutic approach for HIV infection. So far, the use of HAART offers the best opportunity for prolonged and maximal viral suppression, and preservation of the immune system upon HIV infection. Still, the frequent administration of high doses of multiple drugs, their inefficient ability to reach the viral reservoirs in adequate doses, the development of drug resistance, and the lack of patient compliance compromise the complete HIV elimination. The development of nanotechnology-based drug delivery systems may enable targeted delivery of antiretroviral agents to inaccessible viral reservoir sites at therapeutic concentrations. In addition, the application of Computer-Aided Drug Design (CADD) approaches has provided valuable tools for the development of anti-HIV drug candidates with favourable pharmacodynamics and pharmacokinetic properties.

Structural Characterization of Membrane Protein Dimers.

Membrane proteins are essential vessels for cell communication both with other cells and noncellular structures. They modulate environment responses and mediate a myriad of biological processes. Dimerization and multimerization processes have been shown to further increase the already high specificity of these processes. Due to their central role in various cell and organism functions, these multimers are often associated with health conditions, such as Alzheimer's disease (AD), Parkinson's disease (PD), and diabetes, among others.Understanding the membrane protein dimers' interface takes advantage of the specificity of the structure, for which we must pinpoint the most relevant interfacial residues, since they are extremely likely to be crucial for complex formation. Here, we describe step by step our own in silico protocol to characterize these residues, making use of known experimental structures. We detail the computational pipeline from data acquisition and pre-processing to feature extraction. A molecular dynamics simulation protocol to further study membrane dimer proteins and their interfaces is also illustrated.

Treatment adoption and relative effectiveness of aromatase inhibitors compared to tamoxifen in early breast cancer: A multi-institutional observational study.

BACKGROUND: Since 2005, aromatase inhibitors (AIs) have been the adjuvant treatment of choice for postmenopausal women with early breast cancer (BC). In this study we characterize the adoption of AIs in Portugal, variables associated with treatment administration, and compare its effectiveness (either in monotherapy or sequential therapy) to tamoxifen monotherapy (TAM). PATIENTS AND METHODS: This was a retrospective cohort study that included postmenopausal women with stage I-III hormone receptor (HR) positive BC diagnosed from 2006 to 2008 and treated with adjuvant endocrine therapy in four participating institutions. RESULTS: Of the 1283 eligible patients, 527 (41%) received an AI (16% as monotherapy, 25% as sequential therapy) and 756 (59%) TAM. Patients treated with AI had less differentiated tumors, with higher TNM stage, and were more frequently HER2-positive. Use of AI also differed by center (use range from 33% to 75%, p < 0.001). With a median follow-up of 6.3 years and controlling for clinicopathological and treatment characteristics, treatment with AI had a better overall survival (OS) when compared with TAM (adjusted-HR 0.55, 95% CI 0.37-0.81). CONCLUSION: AIs were successfully introduced as adjuvant treatment for HR-positive BC in Portuguese hospitals. Its use was influenced by tumor and patient characteristics, but also center of care. In this large cohort, AI use was associated with an OS benefit.

Variation in type of adjuvant chemotherapy received among patients with stage I breast cancer: A multi-institutional Portuguese cohort study.

BACKGROUND: A contemporary US study showed an increase in the use of chemotherapy in the last decade for some patients with stage-I breast cancer; with a rise in more intensive regimens, and declining use of anthracyclines. Nevertheless, there is still uncertainty on the absolute benefit of chemotherapy for these patients and the optimal regimen. In this study we compare those findings with the patterns of care among a Portuguese cohort of stage-I breast cancers. METHODS: Retrospective cohort study of patients with stage-I breast cancer diagnosed from 2006 to 2008 at four Portuguese institutions. The use and type of chemotherapy was evaluated. RESULTS: Among patients with stage I-III breast cancer 39.4% (n = 682) had stage I disease. Of the 595 eligible patients, 22.4% were treated with chemotherapy, 33.9% aged <55 years vs. 12.7% aged >65 years (p < 0.001). Thirteen percent of patients with hormone receptor (HR)+/HER2- tumors, 52.7% of patients with HER2+ and 66.0% of patients with HR-/HER2- received chemotherapy (p < 0.001). In addition, we found inter-institutional variability, with the use of chemotherapy ranging from 0.0% to 43.4% (p < 0.001). Eighty-five percent of patients treated with chemotherapy received less-intensive regimens with anthracycline-based regimens, such as doxorubicin and cyclophosphamide, being the most frequently used, while docetaxel and cyclophosphamide was only used in 1.5% of cases. CONCLUSIONS: Overall, almost one-quarter of patients received chemotherapy with institutional variability. When treated, mostly less-intensive associations including anthracyclines were used, which contrasts with contemporary US practice. This study highlights the need for health-services research to understand local practices and tailor quality improvement interventions.