26 April 2021 in IEEE Transactions on Network and Service Management
Network Intrusion Detection Systems (NIDS) play an important role as tools for identifying potential network threats. In the context of ever-increasing traffic volume on computer networks, flow-based NIDS arise as good solutions for real-time traffic classification. In recent years, different flow-based classifiers have been proposed using Machine Learning (ML) algorithms. Nevertheless, classical ML-based classifiers have some limitations. For instance, they require large amounts of labeled data for training, which might be difficult to obtain. Additionally, most ML-based classifiers are not capable of domain adaptation, i.e., after being trained on an specific data distribution, they are not general enough to be applied to other related data distributions. And, finally, many of the models inferred by these algorithms are black boxes, which do not provide explainable results. To overcome these limitations, we propose a new algorithm, called Energy-based Flow Classifier (EFC). This anomaly-based classifier uses inverse statistics to infer a statistical model based on labeled benign examples. We show that EFC is capable of accurately performing binary flow classification and is more adaptable to different data distributions than classical ML-based classifiers. Given the positive results obtained on three different datasets (CIDDS-001, CICIDS17 and CICDDoS19), we consider EFC to be a promising algorithm to perform robust flow-based traffic classification.
16 December 2023 in Nature Communications
Energetic local frustration offers a biophysical perspective to interpret the effects of sequence variability on protein families. Here we present a methodology to analyze local frustration patterns within protein families and superfamilies that allows us to uncover constraints related to stability and function, and identify differential frustration patterns in families with a common ancestry. We analyze these signals in very well studied protein families such as PDZ, SH3, ɑ and β globins and RAS families. Recent advances in protein structure prediction make it possible to analyze a vast majority of the protein space. An automatic and unsupervised proteome-wide analysis on the SARS-CoV-2 virus demonstrates the potential of our approach to enhance our understanding of the natural phenotypic diversity of protein families beyond single protein instances. We apply our method to modify biophysical properties of natural proteins based on their family properties, as well as perform unsupervised analysis of large datasets to shed light on the physicochemical signatures of poorly characterized proteins such as the ones belonging to emergent pathogens.
3 November 2021 in IEEE Transactions on Network and Service Management
Abnormal behavior in mobile cellular networks can cause network faults and consequent cell outages, a major reason for operational cost increase and revenue loss for operators. Nonetheless, network faults and cell outages can be avoided by monitoring abnormal situations in the network and acting accordingly. Thus, anomaly detection is an important component of self-healing control and network management. Network operators may use the detected abnormal behavior to quantify numerically their intensity. The quantification of abnormal behavior assists the characterization of potential regions for infrastructure updates and to support the creation of public policies for local connectivity enhancements. We propose an unsupervised learning solution for anomaly detection in mobile networks using Call Detail Records (CDR) data. We evaluate our solution using a real CDR data set provided by an Italian operator and compare it against other state-of-the-art solutions, showing a performance improvement of around 35%. We also demonstrate the relevance of considering the distinct traffic patterns of diverging geographic areas for anomaly detection in mobile networks, an aspect often ignored in the literature.
3 October 2021 in Proteins: Structure, Function, and Bioinformatics
We present the results of the assessment of the intramolecular residue–residue contact and distance predictions from groups participating in the 14th round of the CASP experiment. The performance of contact prediction methods was evaluated with the measures used in previous CASPs, while distance predictions were assessed based on a new protocol, which considers individual distance pairs as well as the whole predicted distance matrix, using a graph‐based framework. The results of the evaluation indicate that predictions by the tFold framework, TripletRes and DeepPotential were the most accurate in both categories. With regards to progress in method performance, the results of the assessment in contact prediction did not reveal any discernible difference when compared to CASP13. Arguably, this could be due to CASP14 FM targets being more challenging than ever before.
21 August 2020 in bioRxiv
The recent emergence of the novel SARS-CoV-2 in China and its rapid spread in the human population has led to a public health crisis worldwide. Like in SARS-CoV, horseshoe bats currently represent the most likely candidate animal source for SARS-CoV-2. Yet, the specific mechanisms of cross-species transmission and adaptation to the human host remain unknown. Here we show that the unsupervised analysis of conservation patterns across the β-CoV spike protein family, using sequence information alone, can provide rich information on the molecular basis of the specificity of β-CoVs to different host cell receptors. More precisely, our results indicate that host cell receptor usage is encoded in the amino acid sequences of different CoV spike proteins in the form of a set of specificity determining positions (SDPs). Furthermore, by integrating structural data, in silico mutagenesis and coevolution analysis we could elucidate the role of SDPs in mediating ACE2 binding across the Sarbecovirus lineage, either by engaging the receptor through direct intermolecular interactions or by affecting the local environment of the receptor binding motif. Finally, by the analysis of coevolving mutations across a paired MSA we were able to identify key intermolecular contacts occurring at the spike-ACE2 interface. These results show that effective mining of the evolutionary records held in the sequence of the spike protein family can help tracing the molecular mechanisms behind the evolution and host-receptors adaptation of circulating and future novel β-CoVs. Significance: Unraveling the molecular basis for host cell receptor usage among β-CoVs is crucial to our understanding of cross-species transmission, adaptation and for molecular-guided epidemiological monitoring of potential outbreaks. In the present study, we survey the sequence conservation patterns of the β-CoV spike protein family to identify the evolutionary constraints shaping the functional specificity of the protein across the β-CoV lineage. We show that the unsupervised analysis of statistical patterns in a MSA of the spike protein family can help tracing the amino acid space encoding the specificity of β-CoVs to their cognate host cell receptors. We argue that the results obtained in this work can provide a framework for monitoring the evolution of SARS-CoV-2 specificity to the hACE2 receptor, as the virus continues spreading in the human population and differential virulence starts to arise.
27 December 2024 in bioRxiv
MFMiguel Fernandez-Martin
Evolutionarily conserved protein-protein interactions (PPIs) reveal fundamental biological processes. However, predicting protein function solely from these interactions provides an incomplete picture of biological systems. Computational methods for predicting PPIs often struggle due to limited functional annotations in databases, making it difficult to fully understand unique biological systems.. This study introduces ContextMirror2.0 (CM2.0), a coevolution-based method designed to address these limitations. Coevolutionary approaches, unlike supervised machine learning methods, do not rely on labelled datasets, proving valuable in addressing the scarcity of data for species-specific PPIs. While around 40% of CM2.0’s top-1000 predicted PPIs for E. coli, S. enterica , and S. aureus are present in experimental PPI databases, a comparative analysis of predicted functional communities revealed highly conserved interaction patterns and species-specific interactions. CM2.0 leverages coevolutionary information to explore protein interaction dynamics and understand the functional consequences of species-specific variations. This information can inform further studies on the emergence of novel functions, adaptation to specific environments, and the development of targeted therapies.