Avalanche Of Data Research Articles

Deep learning-based models for preimplantation mouse and human embryos based on single-cell RNA sequencing.

The rapid growth of single-cell transcriptomic technology has produced an increasing number of datasets for both embryonic development and in vitro pluripotent stem cell-derived models. This avalanche of data surrounding pluripotency and the process of lineage specification has meant it has become increasingly difficult to define specific cell types or states in vivo, and compare these with in vitro differentiation. Here we utilize a set of deep learning tools to integrate and classify multiple datasets. This allows the definition of both mouse and human embryo cell types, lineages and states, thereby maximizing the information one can garner from these precious experimental resources. Our approaches are built on recent initiatives for large-scale human organ atlases, but here we focus on material that is difficult to obtain and process, spanning early mouse and human development. Using publicly available data for these stages, we test different deep learning approaches and develop a model to classify cell types in an unbiased fashion at the same time as defining the set of genes used by the model to identify lineages, cell types and states. We used our models trained on in vivo development to classify pluripotent stem cell models for both mouse and human development, showcasing the importance of this resource as a dynamic reference for early embryogenesis.

Recommender Systems Applications: Data Sources, Features, and Challenges

In recent years, there has been growing interest in recommendation systems, which is matched by their widespread adoption across various sectors. This can be attributed to their effectiveness in reducing an avalanche of data into individualized information that is meaningful, relevant, and can easily be absorbed by a single person. Several studies have recently navigated the landscape of recommendation systems, attending to their approaches, challenges, and applications, as well as the evaluation metrics necessary for effective implementation. This systematic review investigates the understudied aspects of recommendation systems, including the data input into the systems and their features or outputs. The data in (input) and data out (features) are both diverse and vary significantly from not just one application domain to another, but also from one application use case to another, which is a distinction that has not been thoroughly addressed in the past. In addition, this study explores several application domains, providing a comprehensive breakdown of the categorical data consumed by these systems and the features, or outputs, of these systems. Without focusing on any particular journals or their rankings, this study collects and reviews articles on recommendation systems published from 2018 to April 2024, in four top-tier research repositories, including IEEE Xplore Digital Library, Springer Link, ACM Digital Library, and Google Scholar.

MarkerScan: Separation and assembly of cobionts sequenced alongside target species in biodiversity genomics projects.

Contamination of public databases by mislabelled sequences has been highlighted for many years and the avalanche of novel sequencing data now being deposited has the potential to make databases difficult to use effectively. It is therefore crucial that sequencing projects and database curators perform pre-submission checks to remove obvious contamination and avoid propagating erroneous taxonomic relationships. However, it is important also to recognise that biological contamination of a target sample with unexpected species' DNA can also lead to the discovery of fascinating biological phenomena through the identification of environmental organisms or endosymbionts. Here, we present a novel, integrated method for detection and generation of high-quality genomes of all non-target genomes co-sequenced in eukaryotic genome sequencing projects. After performing taxonomic profiling of an assembly from the raw data, and leveraging the identity of small rRNA sequences discovered therein as markers, a targeted classification approach retrieves and assembles high-quality genomes. The genomes of these cobionts are then not only removed from the target species' genome but also available for further interrogation. Source code is available from https://github.com/CobiontID/MarkerScan. MarkerScan is written in Python and is deployed as a Docker container.

Novel Computational Methods for Cancer Drug Design.

Cancer is a complex and debilitating disease that is one of the leading causes of death in the modern world. Computational methods have contributed to the successful design and development of several drugs. The recent advances in computational methodology, coupled with the avalanche of data being acquired through high throughput genomics, proteomics, and metabolomics, are likely to increase the contribution of computational methods toward the development of more effective treatments for cancer. Recent advances in the application of neural networks for the prediction of the native conformation of proteins have provided structural information regarding the complete human proteome. In addition, advances in machine learning and network pharmacology have provided novel methods for target identification and for the utilization of biological, pharmacological, and clinical databases for the design and development of drugs. This is a review of the key advances in computational methods that have the potential for application in the design and development of drugs for cancer.

Artificial Intelligence Techniques for Uncovering Resolved Planetary Nebula Candidates from Wide-field VPHAS+ Survey Data

ABSTRACT Artificial intelligence (AI) and deep learning techniques are playing an increasing role in astronomy to deal with the data avalanche. Here we describe an application for finding resolved planetary nebulae (PNe) in crowded, wide-field, narrow-band Hα survey imagery in the Galactic plane, to test and facilitate more objective, reproducible, efficient and reliable trawls for them. PNe are important for studying the late-stage stellar evolution of low-mass to intermediate-mass stars. However, the confirmed ∼3800 Galactic PNe fall far short of the numbers expected. Traditional visual searching for resolved PNe is time-consuming because of the large data size and areal coverage of modern astronomical surveys. The training and validation data set of our algorithm was built with the INT Photometric Hα Survey (IPHAS) and true PNe from the Hong Kong/AAO/Strasbourg Hα (HASH) data base. Our algorithm correctly identified 444 PNe in the validation set of 454 PNe, with only 16 explicable ‘false’ positives, achieving a precision rate of 96.5 per cent and a recall rate of 97.8 per cent. After transfer learning, it was then applied to the VST Photometric Hα Survey of the Southern Galactic plane and bulge (VPHAS+), examining 979 out of 2284 survey fields, each covering 1° × 1°. It returned ∼20 000 detections, including 2637 known PNe and other kinds of catalogued non-PNe. A total of 815 new high-quality PNe candidates were found, 31 of which were selected as top-quality targets for optical spectroscopic follow-up. We found that 74 per cent of them are true, likely, and possible PNe. Representative preliminary confirmatory spectroscopy results are presented here to demonstrate the effectiveness of our techniques, with full details to be given in our forthcoming paper.

3 Overview of Poultry Modeling Evolution

Abstract This presentation aims to review advances in mathematical models used in poultry nutrition and production in the past decades and present an overview of the current status. Models have been developed to improve understanding of growth, voluntary feed intake, nutrient utilization, body composition, responses to nutrient supply, environmental changes, and stressors. Growth models for broilers, turkeys, ducks, layers, and breeders are available mainly based on the Gompertz curve, but a few other mathematical functions are also used. A few models can simulate egg production in layers and breeders. The EFG and Avinesp models, tools from feed additive and genetic companies, among others, will be described or mentioned. These models can describe genetic differences and estimate energy, amino acids, calcium, and phosphorus requirements or multiple biological responses to specific nutrient levels. Most models use the concept of ideal protein or balanced protein to describe responses to amino acid levels. However, very few models can describe responses to modifications of one or two specific amino acids. Compartmental models are mainly based on main chemical components such as water, body and feather protein, fat, and ash. Poultry models seldom consider variability due to feeding ingredient quality, fiber, fat, other mineral composition, or variation in nutrient digestibility. However, some models can predict responses to exogenous enzymes. Most applications can predict body weights and yields at a specific age, and some can estimate nitrogen, calcium, and phosphorus excretion. These models can be integrated with feed formulation software, economic data, and optimizers to simulate multiple conditions and estimate biological, economic, and environmental optimums. Poultry enterprise models can describe the whole production system, estimate the optimum combination of resources, aid in planning and scheduling, and determine strategies for growth. Currently, no official recommendations from academic or private groups are based only on models to estimate nutrient levels. Combinations of empirical data and model results are frequently used for nutritional decision-making. A few tools are available for specialized personnel to decide the energy and amino acid levels to use. The application of electronics and telecommunications triggered the fast production of enormous volumes of data from diverse poultry production sectors. This big data avalanche is starting to help modelers to integrate more accurate information and incorporate natural variability. However, integration between current models and significant data results is still under development. Despite increasing international collaboration in modeling, poultry model developments keep great independence and fundamental differences. One critical issue needed for faster development of poultry modeling is to increase education in modeling. The lack of academic programs offering modeling classes related to poultry caused the low current application of models. However, awareness about the benefits of modeling has increased in the past few years, and efforts are under development.

Big Data Analytics: Unveiling Insights and Opportunities

Information technology today is increasingly concerned with dealing with massive data sets. The proliferation of the internet and, by extension, the digital economy has resulted in a meteoric rise in the need for data storage and analysis. This creates a serious problem for American IT departments in terms of securing and analysing the resulting avalanche of data. Businesses currently acquire and store more data than ever before due to the critical role that information plays in their daily operations. In all likelihood, this pattern will maintain its current trajectory. The organised knowledge that is being developed right now is based on a lot of legacy information. Instead, it's information like text, images, music, video, and social media posts. It's called "unstructured knowledge" when the knowledge isn't in any particular shape. The term "big data analytics" refers to a technique that can be used to get insight from these massive datasets. In addition to generating new business prospects, this strategy has been shown to increase the percentage of returning customers.

Hereditable variants of classical protein tyrosine phosphatase genes: Will they prove innocent or guilty?

Protein tyrosine phosphatases, together with protein tyrosine kinases, control many molecular signaling steps that control life at cellular and organismal levels. Impairing alterations in the genes encoding the involved proteins is expected to profoundly affect the quality of life-if compatible with life at all. Here, we review the current knowledge on the effects of germline variants that have been reported for genes encoding a subset of the protein tyrosine phosphatase superfamily; that of the thirty seven classical members. The conclusion must be that the newest genome research tools produced an avalanche of data that suggest 'guilt by association' for individual genes to specific disorders. Future research should face the challenge to investigate these accusations thoroughly and convincingly, to reach a mature genotype-phenotype map for this intriguing protein family.

Building Trust in Medical Use of Artificial Intelligence – The Swarm Learning Principle

ABSTRACTAn avalanche of medical data is starting to be build up. With the digitalisation of medicine and novel approaches such as the omics technologies, we are conquering ever bigger data spaces to be used to describe pathophysiology of diseases, define biomarkers for diagnostic purposes or identify novel drug targets. Utilising this growing lake of medical data will only be possible, if we make use of machine learning, in particular artificial intelligence (AI)-based algorithms. While the technological developments and chances of the data and information sciences are enormous, the use of AI in medicine also bears challenges and many of the current information technologies (IT) do not follow established medical traditions of mentoring, learning together, sharing insights, while preserving patient’s data privacy by patient physician privilege. Other challenges to the medical sector are demands from the scientific community such as “Open Science”, “Open Data”, “Open Access” principles. A major question to be solved is how to guide technological developments in the IT sector to serve well-established medical traditions and processes, yet allow medicine to benefit from the many advantages of state-of-the-art IT. Here, I provide the Swarm Learning (SL) principle as a conceptual framework designed to foster medical standards, processes and traditions. A major difference to current IT solutions is the inherent property of SL to appreciate and acknowledge existing regulations in medicine that have been proven beneficial for patients and medical personal alike for centuries.

Improving efficiency of fitting Cox proportional hazards models for time-to-event outcomes in genome-wide association studies (GWAS).

Technologies identifying single nucleotide polymorphisms (SNPs) in DNA sequencing yield an avalanche of data requiring analysis and interpretation. Standard methods may require many weeks of processing time. The use of statistical methods requiring data sorting, matrix inversions of a high-dimension and replication in subsets of the data on multiple outcomes exacerbate these times.A method which reduces the computational time in problems with time-to-event outcomes and hundreds of thousands/millions of SNPs using Cox-Snell residuals after fitting the Cox proportional hazards model (PH) to a fixed set of concomitant variables is proposed. This yields coefficients for SNP effect from a Cox-Snell adjusted Poisson model and shows a high concordance to the adjusted PH model.The method is illustrated with a sample of 10000 SNPs from a genome-wide association study in a diabetic population. The gain in processing efficiency using the proposed method based on Poisson modelling can be as high as 62%. This could result in saving of over three weeks processing time if 5 million SNPs require analysis. The method involves only a single predictor variable (SNP), offering a simpler, computationally more stable approach to examining and identifying SNP patterns associated with the outcome(s) allowing for a faster development of genetic signatures. Use of deviance residuals from the PH model to screen SNPs demonstrates a large discordance rate at a 0.2% threshold of concordance. This rate is 15 times larger than that based on the Cox-Snell residuals from the Cox-Snell adjusted Poisson model. The method is simple to implement as the procedures are available in most statistical packges. The approach involves obtaining Cox-Snell residuals from a PH model, to a binary time-to-event outcome, for factors which need to be common when assessing each SNP. Each SNP is then fitted as a predictor to the outcome of interest using a Poisson model with the Cox-Snell as the exposure variable.

DeePromClass: Delineator for Eukaryotic Core Promoters Employing Deep Neural Networks.

Computational promoter identification in eukaryotes is a classical biological problem that should be refurbished with the availability of an avalanche of experimental data and emerging deep learning technologies. The current knowledge indicates that eukaryotic core promoters display multifarious signals such as TATA-Box, Inr element, TCT, and Pause-button, etc., and structural motifs such as G-quadruplexes. In the present study, we combined the power of deep learning with a plethora of promoter motifs to delineate promoter and non-promoters gleaned from the statistical properties of DNA sequence arrangement. To this end, we implemented convolutional neural network (CNN) and long short-term memory (LSTM) recurrent neural network architecture for five model systems with [-100 to +50] segments relative to the transcription start site being the core promoter. Unlike previous state-of-the-art tools, which furnish a binary decision of promoter or non-promoter, we classify a chunk of 151mer sequence into a promoter along with the consensus signal type or a non-promoter. The combined CNN-LSTM model; we call "DeePromClass", achieved testing accuracy of 90.6%, 93.6%, 91.8%, 86.5%, and 84.0% for S. cerevisiae, C. elegans, D. melanogaster, Mus musculus, and Homo sapiens respectively. In total, our tool provides an insightful update on next-generation promoter prediction tools for promoter biologists.

Machine learning algorithms: their applications in plant omics and agronomic traits’ improvement

Agronomic traits of plants especially those of economic or aesthetic importance are threatened by climatic and environmental factors such as climate change, biotic, and abiotic stresses. These threats are now being mitigated through the analyses of omics data like genomics, transcriptomics, proteomics, metabolomics, and phenomics. The emergence of high-throughput omics technology has led to an avalanche of plant omics data. Plant research demands novel analytical paradigms to extract and harness large plant omics data for plant improvement effectively and efficiently. Machine learning algorithms are well-suited analytical and computational approaches for the integrative analysis of large unstructured, heterogeneous datasets. This study presents an overview of omics approaches to improve plant agronomic traits and crucial curated plant genomic data sources. Furthermore, we summarize machine learning algorithms and software tools/programming packages used in plant omics research. Lastly, we discuss advancements in machine learning algorithms' applications in improving agronomic traits of economically important plants. Extensive application of machine learning would advance plant omics studies. These advancements would consequently help agricultural scientists improve economically important plants’ quality, yield, and tolerance against abiotic and biotic stresses and other plant health-threatening issues.

Brain-inspired computing needs a master plan

New computing technologies inspired by the brain promise fundamentally different ways to process information with extreme energy efficiency and the ability to handle the avalanche of unstructured and noisy data that we are generating at an ever-increasing rate. To realize this promise requires a brave and coordinated plan to bring together disparate research communities and to provide them with the funding, focus and support needed. We have done this in the past with digital technologies; we are in the process of doing it with quantum technologies; can we now do it for brain-inspired computing?

Brain Asymmetry: Towards an Asymmetrical Neurovisceral Integration

Despite the ancestral evidence of an asymmetry in motor predominance, going through the inspiring discoveries of Broca and Wernicke on the localization of language processing, continuing with the subsequent noise coinciding with the study of brain function in commissurotomized patients—and the subsequent avalanche of data on the asymmetric distribution of multiple types of neurotransmitters in physiological and pathological conditions—even today, the functional significance of brain asymmetry is still unknown. Currently, multiple evidence suggests that functional asymmetries must have a neurochemical substrate and that brain asymmetry is not a static concept but rather a dynamic one, with intra- and inter-hemispheric interactions between its various processes, and that it is modifiable depending on changing endogenous and environmental conditions. Furthermore, based on the concept of neurovisceral integration in the overall functioning of an organism, some evidence has emerged suggesting that this integration could be organized asymmetrically, using the autonomic nervous system as a bidirectional communication pathway, whose performance would also be asymmetric. However, the functional significance of this distribution, as well as the evolutionary advantage of an asymmetric nervous organization, is still unknown.

Influence of Big Data Technology on Enterprise Marketing Strategy

In today's society, with the rapid development of Internet technology, we are faced with a veritable avalanche of different data. This effectively symbolizes the advent of the big data era. What new opportunities and challenges can this availability of information bring to our lives? How will the formulation of marketing strategies of related enterprises be affected and changed at the advent of this new era? In today's fast-growing Internet environment, how to entice enterprises to make rational use of the advantages available in big data, and how to guide enterprises to formulate and plan important marketing strategies are just two of the various practical problems that all entrepreneurs should be considering. This paper will focus on analyzing the actual impact of massive data and information on an enterprise's marketing strategy. At the same time, we should also put forward the optimal thinking with regard to guiding marketing strategies of enterprises through big data, according to the actual situation of those enterprises. The main purpose of this study is to actually help the relevant enterprises improve their ability to formulate marketing strategies in this new era of big data, At the same time, enterprises should learn to use big data as a means to influence and create the new advantages and driving forces that are conducive to enterprise development.

Image Augmentation based on GAN deep learning approach with Textual Content Descriptors

Computer vision, also known as computational visual perception, is a branch of artificial intelligence that allows computers to interpret digital pictures and videos in a manner comparable to biological vision. It entails the development of techniques for simulating biological vision. The aim of computer vision is to extract more meaningful information from visual input than that of a biological vision. Computer vision is exploding due to the avalanche of data being produced today. Powerful generative models, such as Generative Adversarial Networks (GANs), are responsible for significant advances in the field of picture creation. The focus of this research is to concentrate on textual content descriptors in the images used by GANs to generate synthetic data from the MNIST dataset to either supplement or replace the original data while training classifiers. This can provide better performance than other traditional image enlarging procedures due to the good handling of synthetic data. It shows that training classifiers on synthetic data are as effective as training them on pure data alone, and it also reveals that, for small training data sets, supplementing the dataset by first training GANs on the data may lead to a significant increase in classifier performance.

Data Science for Official Statistics: Views of the United Nations Statistics Division

In the digital age, data are generated continuously by many different devices and are being used by many different actors. National statistical offices (NSOs) should benefit from these opportunities to improve data for decision-making. What could be the expanding role for official statistics in this context and how does this relate to emerging disciplines like data science? This article explores some new ideas. In the avalanche of new data, society may need a data steward, and the NSO could take on that role, while paying close attention to the protection of privacy. Data science will become increasingly important for extracting meaningful information from large amounts of data. NSOs will need to hire data scientists and data engineers and will need to train their staff in these fast-developing fields. NSOs will also need to clearly communicate new and experimental data and foster a good understanding of statistics. Collaboration of official statistics with the private sector, academia, and civil society will be the new way of working and the fundamental principles of official statistics may have to apply to all those actors. This article envisions that we are gradually working toward such a new data culture.

Transfer RNA and Origins of RNA Interference.

Almost 30 years ago the first microRNA (miRNA) was detected (Lee et al., 1993), later put in mechanistical context by the discovery of the RNA interference (RNAi) and RNA-induced silencing complex (RISC). miRNA is the variable RISC element (guide) that recognizes messenger RNA (mRNA) targets in the RNAi process facilitated by Argonaute (Ago) proteins, the fixed key players in RISC. Alas, poetically sounding Argonautes were named not after the adventurers from Greek mythology but octopus-shaped leaves of an Arabidopsis thaliana mutant (Bohmert et al., 1998). Production of miRNAs is performed by a sophisticated orchestra of players with even groovier names, like Drosha, Pasha or Dicer. While roles of miRNA in post-transcriptional regulation have been established, its evolution has been typically described as expansion of new miRNA genes from the existing ones, not their first appearance (Berezikov, 2011). The emergence of RNAi has been attributed to its defense against pathogens, which involves other participants, e.g., double-stranded RNA binding proteins (dsRBPs). But details of such emergence are lacking. This is a short opinion piece, not a full review, so I apologetically skip some of these players and hundreds of relevant miRNA/RNAi citations (and those for subsections below). The complexity of the RISC is bewildering. Intricate steps of miRNA production in the nucleus are followed by unequal strand loading to Ago, and finally resulting in cytoplasmic events, sometimes involving “slashing” of mRNA. All that (for simplicity, ignoring nuclear events upon Ago transport from cytoplasm) is based on weak and imperfect binding of short “seeds” (5′ 6-8 nucleotides in length in animals) to targets and on weak outcomes of slight tuning of target translation. How can such weak binding be a driving force for emergence and evolution of this complicated system and its parts, acting in concert in different cellular compartments? Plant RNAi requires full-length small RNA hybridization, significantly limiting the number of target genes, yet the system appears even more complex (four Dicers, seven dsRBPs versus a single Dicer/couple of dsRBPs in humans). Further, the miRNA generation machineries are not entirely homologous, mature miRNAs are produced and modified in the plant nucleus, but these processes are divided between the nucleus and cytoplasm in animals. How would these different processes evolve in parallel with creating the elaborate and divergent Ago functionalities (four different Agos in humans, ten in A. thaliana, 26 in Caenorhabditis elegans), able to utilize these precisely cut miRNAs for mRNA regulation? This scenario would require a small RNA already acting together with some Ago prototypic protein. Compatible with this view, the last common ancestor of eukaryotes likely appeared after the RNAi functional principles (and some components, except for miRNA) had been invented (Cerutti and Casas-Mollano, 2006). And many prokaryotes do not have miRNAs but possess Ago protein homologs. Transfer RNAs and Their Fragments Enter transfer RNAs (tRNAs), fundamental elements in mRNA translation. Each tRNA is a link of informational (anticodon) and corresponding chemical (aminoacid) units, which jointly provide the basis of the central dogma. These are ancient molecules, potentially capable of performing primitive replication (Kuhnlein et al., 2021) or aiding in it (Maizels and Weiner, 1994). While the genetic code requires <64 anticodons for translation, tRNA genes are very numerous, with several hundred copies in human genome. From a regulation standpoint, these numbers far exceed potential codon adaptation mechanisms. Perhaps fitting to this view, many tRNA genes have been described as inactive, raising questions about their actual role (Torres, 2019). The early sequence/structure determination and clearly defined function of tRNAs established their place in molecular biology textbooks, hardly revisited until the arrival of tRNA fragments (tRFs). Hypotheses of further roles were occasionally entertained, based on additional function observed, e.g., in viral replication, etc. (see also a review (Avcilar-Kucukgoze and Kashina, 2020) in this article collection for more details on these roles). However, an avalanche of data from small RNA sequencing experiments have challenged such perceptions, revealing numerous and ubiquitous tRFs in a multitude of sequenced samples. These fragments, detected in datasets produced to study miRNA, have been mostly dismissed as noise. It took >10 years to get acceptance even after detailed studies (Cole et al., 2009; Lee et al., 2009; Haussecker et al., 2010; Ivanov et al., 2011; Gebetsberger et al., 2012), even though tRNA breakage products were detected in urine of cancer patients much earlier (Speer et al., 1979). The finding of Ago proteins loaded with tRFs, in addition to their cargo of miRNAs, have prompted speculations about similar functionality of tRFs and search for their targets (Table 1). TABLE 1 Representative studies on detection and analysis of Argonaute-loaded tRFs in different organisms.

Improving the Development Technology of an Oil and Gas Company Using the Minimax Optimality Criterion

The article deals with the problem of adaptation of the Russian oil and gas company (Novatek, Russia) to the rapidly changing external environment, the avalanche of data from competitors, and the need to filter important information for business development and the prosperity of the industry as a whole. The approach is based on the system of integrated software monitoring of key business processes at the enterprise developed by the authors—from the formation of the idea of a new product to its implementation to paying customers. The scientific novelty lies in the use of an optimization model that allows for minimizing the maximum losses of the investor at all levels of decision-making, from the distribution of capital between companies, to the optimization of internal reserves to increase the competitiveness of the company. The toolkit is a minimax model that allows you to redistribute the shares of investor influence at the portfolio level, and then within the business processes of each company selected by investors, in order to achieve the optimal solution in accordance with the selected estimated indicators. Application of the well-known portfolio investment models of Markowitz, Tobin, Sharp, etc. is not possible due to the lack of necessary data on the basis of which the probabilistic parameters involved in the model are estimated. Even if we get them, it is necessary to take into account the level of correlation influence of the technological process in the composition of each subsystem, which is unacceptable for the data used, as it leads to a strong increase in errors. Using minimax and a systematic approach allows you to minimize such errors by choosing a balanced concentration of distributed assets for both the investor and the buyer. To this end, a three-way analysis of the company’s development was carried out and a technology for comprehensive improvement of the company’s activities was developed in the following areas: the company’s rating in the industry, financial condition, and interaction with counterparties using merchandising technologies. Tools for optimal image zoning at the Novatek site using the minimax approximation criterion have been developed. The technology provides a procedure for creating a comfortable mode of image perception based on high-tech visualization of merchandising, zoning of the screen area, and a mathematical approach that allows you to develop a calculation algorithm.

Exploring parallel MPI fault tolerance mechanisms for phylogenetic inference with RAxML-NG.

MotivationPhylogenetic trees are now routinely inferred on large scale high performance computing systems with thousands of cores as the parallel scalability of phylogenetic inference tools has improved over the past years to cope with the molecular data avalanche. Thus, the parallel fault tolerance of phylogenetic inference tools has become a relevant challenge. To this end, we explore parallel fault tolerance mechanisms and algorithms, the software modifications required and the performance penalties induced via enabling parallel fault tolerance by example of RAxML-NG, the successor of the widely used RAxML tool for maximum likelihood-based phylogenetic tree inference.ResultsWe find that the slowdown induced by the necessary additional recovery mechanisms in RAxML-NG is on average 1.00 ± 0.04. The overall slowdown by using these recovery mechanisms in conjunction with a fault-tolerant Message Passing Interface implementation amounts to on average 1.7 ± 0.6 for large empirical datasets. Via failure simulations, we show that RAxML-NG can successfully recover from multiple simultaneous failures, subsequent failures, failures during recovery and failures during checkpointing. Recoveries are automatic and transparent to the user.Availability and implementationThe modified fault-tolerant RAxML-NG code is available under GNU GPL at https://github.com/lukashuebner/ft-raxml-ng.Supplementary information Supplementary data are available at Bioinformatics online.