Modern commercial antivirus systems increasingly rely on machine learning (ML) to keep up with the rampant inflation of new malware. However, it is well-known that machine learning models are vulnerable to adversarial examples (AEs). Previous works have shown that ML malware classifiers are fragile to the white-box adversarial attacks. However, ML models used in commercial antivirus (AV) products are usually not available to attackers and only return hard classification labels. Therefore, it is more practical to evaluate the robustness of ML models and real-world AVs in a pure black-box manner. We propose a black-box Reinforcement Learning (RL) based framework to generate AEs for PE malware classifiers and AV engines. It regards the adversarial attack problem as a multi-armed bandit problem, which finds an optimal balance between exploiting the successful patterns and exploring more varieties. Compared to other frameworks, our improvements lie in three points: 1) limiting the exploration space by modeling the generation process as a stateless process to avoid combination explosions, 2) reusing the successful payload in modeling; and 3) minimizing the changes on AE samples to correctly assign the rewards in RL learning (which also helps identify the root cause of evasions). As a result, our framework has much higher evasion rates than other off-the-shelf frameworks. Results show it has over 74%--97% evasion rate for two state-of-the-art ML detectors and over 32%--48% evasion rate for commercial AVs in a pure black-box setting. We also demonstrate that the transferability of adversarial attacks among ML-based classifiers is higher than that between ML-based classifiers and commercial AVs.

MOTIVATION: The potentially low precision associated with the geographic origin of sampled sequences represents an important limitation for spatially explicit (i.e. continuous) phylogeographic inference of fast-evolving pathogens such as RNA viruses. A substantial proportion of publicly available sequences is geo-referenced at broad spatial scale such as the administrative unit of origin, rather than more precise locations (e.g. geographic coordinates). Most frequently, such sequences are either discarded prior to continuous phylogeographic inference or arbitrarily assigned to the geographic coordinates of the centroid of their administrative area of origin for lack of a better alternative. RESULTS: We here implement and describe a new approach that allows to incorporate heterogeneous prior sampling probabilities over a geographic area. External data, such as outbreak locations, are used to specify these prior sampling probabilities over a collection of sub-polygons. We apply this new method to the analysis of highly pathogenic avian influenza H5N1 clade data in the Mekong region. Our method allows to properly include, in continuous phylogeographic analyses, H5N1 sequences that are only associated with large administrative areas of origin and assign them with more accurate locations. Finally, we use continuous phylogeographic reconstructions to analyse the dispersal dynamics of different H5N1 clades and investigate the impact of environmental factors on lineage dispersal velocities. AVAILABILITY AND IMPLEMENTATION: Our new method allowing heterogeneous sampling priors for continuous phylogeographic inference is implemented in the open-source multi-platform software package BEAST 1.10. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

The rapid detection and identification of infectious pathogens are a critical need for healthcare in both developed and developing countries. There is a need for portable diagnostics and epidemiological surveillance systems that can be deployed in remote field settings for infectious diseases testing, such as pandemic "hotspots" in Africa or southeast Asia. Together with the Vital-IT and Swiss-Prot groups, we developed a fast pathogen detection embedded hardware accelerator (no bigger than 75x50mm) capable of analyzing up to 250Mnt/s of DNA sequence. To deploy this solution, an embedded DNA Next Generation Sequencer (NGS) is used for retrieving the genetic material from patient's sample at the point of care (e.g., the Oxford Nanopore Technologies MinION). Successively, the collected DNA sequences are analized in order to detect the presence of known pathogen patterns. Currently, up to 100 different pathogens can be analized at the same time using one single embedded accelerator. Due to the low-power consumption of this hardware accelerator, it is possible to give access to worldwide health organizations and local administrations to a complete real-time embedded and battery powered surveillance solution. Moreover, this solution can greatly scale to big computing clusters, where a larger pathogens number can be analized in parallel, outperforming current state of art computationally expensive solutions.

MOTIVATION: Lipids are a large and diverse group of biological molecules with roles in membrane formation, energy storage and signaling. Cellular lipidomes may contain tens of thousands of structures, a staggering degree of complexity whose significance is not yet fully understood. High-throughput mass spectrometry-based platforms provide a means to study this complexity, but the interpretation of lipidomic data and its integration with prior knowledge of lipid biology suffers from a lack of appropriate tools to manage the data and extract knowledge from it. RESULTS: To facilitate the description and exploration of lipidomic data and its integration with prior biological knowledge, we have developed a knowledge resource for lipids and their biology-SwissLipids. SwissLipids provides curated knowledge of lipid structures and metabolism which is used to generate an in silico library of feasible lipid structures. These are arranged in a hierarchical classification that links mass spectrometry analytical outputs to all possible lipid structures, metabolic reactions and enzymes. SwissLipids provides a reference namespace for lipidomic data publication, data exploration and hypothesis generation. The current version of SwissLipids includes over 244 000 known and theoretically possible lipid structures, over 800 proteins, and curated links to published knowledge from over 620 peer-reviewed publications. We are continually updating the SwissLipids hierarchy with new lipid categories and new expert curated knowledge. AVAILABILITY: SwissLipids is freely available at http://www.swisslipids.org/. CONTACT: [email protected] SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.