Non-redundant Data Research Articles

Identifying 124 new anti-HIV drug candidates in a 37 billion-compound database: An integrated approach of machine learning (QSAR), molecular docking, and molecular dynamics simulation

Recent data from the World Health Organization reveals that in 2023, 38.8 million people were living with HIV. Within this population, there were 1.5 million new cases and 650 thousand deaths attributed to the disease. This study employs an integrated approach involving QSAR-based machine learning models, molecular docking, and molecular dynamics simulations to identify potential compounds for inhibiting the bioactivity of the CC chemokine receptor type 5 (CCR5) protein, a key entry point for the HIV virus. Using non-redundant experimental data from the CHEMBL database, 40 different machine learning algorithms were trained and the top four models (XGBoost, Histogram based gradient Boosting, Light Gradient Boosted Machine, and Extra Trees Regression) were utilized to predict anti-HIV bioactivity for 37 billion compounds in the ZINC-22 database. The screening resulted in the identification of 124 new anti-HIV drug candidates, confirmed through molecular docking and dynamics simulations. The study underscores the therapeutic potential of these compounds, paving the way for further in vitro and in vivo investigations. The convergence of machine learning and experimental findings presents a promising avenue for significant advancements in pharmaceutical research, particularly in the treatment of viral diseases such as HIV. To guarantee the reproducibility of our study, we have made the Python code (google colab) and the associated database available on GitHub. You can access them through the following link: GitHub Link: https://github.com/AlexandreCOBRE/code.

Enhancing algorithmic assessment in education: Equi-fused-data-based SMOTE for balanced learning

Recently, there has been a growing interest among researchers in enhancing the efficacy of learning through the utilization of diverse machine learning models within the field of artificial intelligence. However, imbalanced data distributions in educational datasets present a significant challenge to machine learning algorithms. This imbalance can result in biased models, untrustworthy outcomes, and poor performance. Data was gathered from a sample of 2176 first-year novice programming students in this study. Due to an alarming 76% failure rate, the imbalanced dataset was preprocessed before being oversampled with techniques such as SMOTE, SMOTE Borderline, SMOTE-ENN, and ADASYN. The proposed non-redundant synthetic data cooperation approach, named Equi-Fused-Data-based SMOTE, seeks to capitalize on the diversity of the obtained data by combining oversampled datasets. The balanced bagging model was then applied to the combined dataset to demonstrate the robustness of this approach. The promising results demonstrate the effectiveness of the Equi-Fused-Data-based SMOTE model, which achieved a higher Accuracy of 93.85%, a Precision, Recall and F1-score of 92,86%, and an AUC of 98.08%.

FishSNP: a high quality cross-species SNP database of fishes

The progress of aquaculture heavily depends on the efficient utilization of diverse genetic resources to enhance production efficiency and maximize profitability. Single nucleotide polymorphisms (SNPs) have been widely used in the study of aquaculture genomics, genetics, and breeding research since they are the most prevalent molecular markers on the genome. Currently, a large number of SNP markers from cultured fish species are scattered in individual studies, making querying complicated and data reuse problematic. We compiled relevant SNP data from literature and public databases to create a fish SNP database, FishSNP (http://bioinfo.ihb.ac.cn/fishsnp), and also used a unified analysis pipeline to process raw data that the author of the literature did not perform SNP calling on to obtain SNPs with high reliability. This database presently contains 45,690,243 (45 million) nonredundant SNP data for 13 fish species, with 30,288,958 (30 million) of those being high-quality SNPs. The main function of FishSNP is to search, browse, annotate and download SNPs, which provide researchers various and comprehensive associated information.

Single-petal rosette trajectory for 2D functional lung imaging

Purpose: The purpose of this study was to investigate the use of a 2D rosette trajectory for breath-hold and free-breathing real-time imaging of the lungs.Methods: Eight healthy volunteers underwent breath-hold magnetic resonance imaging (MRI) using two different parametrizations of the proposed single-petal rosette (SPR) trajectory, as well as radial ultra-short echo time (UTE) acquisition combined with the tiny golden-angle acquisition scheme. The additional free-breathing acquisitions of all trajectories were performed. The proposed technique was compared with the conventional radial UTE technique regarding image sharpness, signal-to-noise ratio (SNR), regional fractional ventilation (FV) for breath-hold and retrospective image-based self-gating, and real-time imaging capabilities.Results: Image sharpness significantly increased for gated SPR images compared to radial UTE for the end-expiratory stage; no significant difference was found for the self-gated end-inspiratory stage and the breath-hold acquisitions. The SPR trajectory performs significantly better than radial UTE concerning the SNR for breath-hold imaging, whereas no significant difference was found for self-gated images. Fractional ventilation values were comparable between SPR and radial UTE. The SPR real-time data showed a marked reduction in the aliasing artifacts with no apparent streaks.Conclusion: The second half of the petal of the SPR trajectory accomplishes rephasing while acquiring non-redundant k-space data, thus leading to a more efficient sampling of the k-space than that with the radial UTE sampling scheme, with a minimal increase in TR. This leads to the need of less read-outs to achieve the same SNR values and, thus, a reduction in scan time. For real-time application, the SPR approach resulted in a marked reduction of aliasing artifacts when compared to UTE.

Blade profile extraction and edge completion method based on structured light measurement point cloud

Blade characteristic parameters evaluation plays an important role in quality monitoring and processing. In this research, a new blade profile extraction and edge completion strategy is proposed to deal with structured light measurement steam turbine blade. To handle the point cloud discretization and local data missing, the blade cross-sectional profiles are extracted adaptively at different heights, providing accurate and non-redundant data. Using NURBS and circular curves, the missing edge completion combines actual measurement curve trends and blade design geometric constraints, achieving a higher success rate and accuracy. Compared with the Perproj method and Virtual Edge method, the profile extraction accuracy of the proposed method is improved by 65 % and 36 % respectively, and the whole process of a sampled blade with additional noise is simulated to verify the effectiveness. Finally, more than two hundred blades are verified on the structured light measurement system, and the average absolute error of all parameters is less than 0.120 mm, which meets the accuracy requirements of blade quality evaluation in engineering applications.

Learning shared and non-redundant label-specific features for partial multi-label classification

Partial multi-label learning (PML) is designed to address the challenge of having both ground-truth labels and noisy labels in the label set of training instances. In real-world applications, there are often noisy features in addition to noisy labels, but existing PML methods fail to filter out these noisy features or account for feature correlations, label correlations, and feature-label correlations together in the feature learning process. These oversights lead to poor classification accuracy, as the comprehensive exploration of feature and label information in PML can contribute significantly to disambiguation. To address this issue, we propose a novel PML framework that disambiguates candidate labels and learns label-specific features while taking into account all three types of correlations between features and labels. First, we capture high-order label correlations by employing the low-rank representation to obtain a complementary label matrix induced from the partial label matrix, which accounts for the presence of noisy labels. Then, we learn a shared and non-redundant label-specific data representation by incorporating intrinsic feature correlations and the learned label correlations to mitigate the impact of noisy features. Finally, we build a classifier by mapping the label-specific feature representation to the complementary label matrix to model the feature-label correlations. Various experiments validate the superiority of our method.

An Effective Near-Field to Far-Field Transformation with Planar Spiral Scanning for Flat Antennas under Test.

The goal of this article is to provide numerical and experimental assessments of an effective near-field to far-field transformation (NF-FF T) technique with planar spiral scanning for flat antennas under test (AUTs), which requires a non-redundant, i.e., minimum, number of NF measurements. This technique has its roots in the theory of non-redundant sampling representations of electromagnetic fields and was devised by suitably applying the unified theory of spiral scans for non-volumetric antennas to the case in which the considered AUT is modeled by a circular disk having its radius equal to half of the AUT's maximum dimension. It makes use of a 2D optimal sampling interpolation (OSI) formula to accurately determine the massive amount of NF data required by the classical plane-rectangular NF-FF T technique from the non-redundant data gathered along the spiral. It must be emphasized that, when considering flat AUTs, the developed transformation allows one to further and significantly save measurement time as compared to that required by the previously developed NF-FF T techniques with planar spiral scans based on a quasi-planar antenna modeling, because the number of turns of the spiral and that of NF data to be acquired depend somewhat on the area of the modeling surface. The reported numerical simulations assess the accuracy of the proposed NF-FF T technique, whereas the experimental tests prove its practical feasibility.

Entangled Motifs in Membrane Protein Structures.

Entangled motifs are found in one-third of protein domain structures, a reference set that contains mostly globular proteins. Their properties suggest a connection with co-translational folding. Here, we wish to investigate the presence and properties of entangled motifs in membrane protein structures. From existing databases, we build a non-redundant data set of membrane protein domains, annotated with the monotopic/transmembrane and peripheral/integral labels. We evaluate the presence of entangled motifs using the Gaussian entanglement indicator. We find that entangled motifs appear in one-fifth of transmembrane and one-fourth of monotopic proteins. Surprisingly, the main features of the distribution of the values of the entanglement indicator are similar to the reference case of general proteins. The distribution is conserved across different organisms. Differences with respect to the reference set emerge when considering the chirality of entangled motifs. Although the same chirality bias is found for single-winding motifs in both membrane and reference proteins, the bias is reversed, strikingly, for double-winding motifs only in the reference set. We speculate that these observations can be rationalized in terms of the constraints exerted on the nascent chain by the co-translational bio-genesis machinery, which is different for membrane and globular proteins.

A Novel PAPR Reduction in Filter Bank Multi-Carrier (FBMC) with Offset Quadrature Amplitude Modulation (OQAM) Based VLC Systems

The peak to average power ratio (PAPR) is one of the major problem with multicarrier-based systems. Due to its improved spectral efficiency and decreased PAPR, Filter Bank Multicarrier (FBMC) has recently become an effective alternative to the orthogonal multiplexing division (OFDM). For filter bank multicarrier communication/offset quadrature amplitude modulation-Visible light communication (FBMC/OQAM-VLC) systems is proposed a PAPR reduction technique. The suggested approach overlaps the proposed FBMC/OQAM-based VLC data signal with the existing signals. Non-redundant signals and data signals do not overlap in the frequency domain because data signals are scattered on odd subcarriers whereas built signals use even subcarriers. To reduce the effects of large-amplitude signal reduction, the suggested technique converts negative signals into positive signals rather than clipping them off as in conventional FBMC-based VLC systems. The PAPR reduction and bit error rate (BER) are realized using a scaling factor in the transformed signals. Complementary cumulative distribution function(CCDF) and BER are used to calculate the performance of the proposed approach. The presented study found that FBMC/OQAM-VLC systems to achieve a good trade-off between PAPR reduction and BER.

Optimal Sensor Positioning for the Sparse Measurement in the Acoustic Reconstruction of a Large Source

Near-field acoustic holography (NAH) has become an effective tool for acoustic source identification. It usually requires many microphones near the source surface to catch sufficient evanescent waves for good spatial resolution. Besides, such a measurement is time-consuming and uneconomical in practice; standard measurements using densely and uniformly populated sensors are sometimes impossible when dealing with massive sources. Eventually, the compressive sensing (CS) technique has been suggested for NAH, which employs a sparse number of sensors. Most existing works in NAH based on compressive sensing theory have focused on developing acoustic models and algorithms; In contrast, less attention has been paid to the best practice of sensor placement. The present study aims to propose and validate the methods for the optimal selection of sparse measuring points using the compression technique for NAH. Due to the small number of measuring sensors, the transfer matrix between source and hologram points becomes underdetermined. Under such a data condition, even though every piece of data would be felt precious for the reconstruction, the information from some sensors should be discarded further because not all data provide meaningfully independent information. To this end, effective independence and singular-value monitoring methods are used to test the selection of the optimal sensor position from a set of candidate positions. The former chooses non-redundant measurement data, whereas the latter keeps the low condition number of the transfer matrix. Simulation results with vibrating plates show that optimized sensor positions could effectively facilitate the optimality of the approximate solution for CS theory compared to uniformly or randomly distributed measurements. It is because the established transfer matrix has the lowest condition number due to being sparser and more independent sensor positions. Experimental results exhibit a better reconstruction error (10–40%) at 200–2,000 Hz when the sensor positions are optimized than the uniform or random selection approaches.

Distributed Information Exchange With Low Latency for Decision Making in Vehicular Fog Computing

Traditional decision making in a vehicle network includes uploading vehicle sensing data to faraway cloud platforms and then returning correlated results to the vehicles. The data have features of large quantity and high redundancy, which causes high communication latency and vehicle applications to deteriorate. Vehicular fog computing (VFC) is a new network paradigm that uses local fog nodes for decision making. However, how to achieve distributed information exchange with low latency is a challenging issue because the connectivity of the vehicle network is low due to vehicle mobility. In this article, a distributed information exchange scheme with low latency in VFC is proposed. First, considering the frequent changes in vehicle positions and the randomness in driving routes, public transportation facilities with a wider driving range such as buses and taxis are used as fog nodes to increase the probability of uploading data. Then, the fog nodes should dynamically adjust the data sampling frequency according to the time-space correlation of the data to ensure that only nonredundant data are received. To minimize the interruption latency caused by accidents during an exchange, the fog nodes evaluate and predict connection states among them and their neighboring vehicles when establishing exchanges. If a fog node finds that a vehicle cannot complete information exchange because the vehicle may move outside its communication range in a future period, it will recalculate an optimized relay route for the vehicle by using mixed integer programming. Theoretical analysis and simulation results show that compared with the existing work, the proposed scheme can completely exchange all vehicle data with lower latency.

Millisecond autofocusing microscopy using neuromorphic event sensing

Rapid autofocusing is essential for many microscopic imaging applications. Existing methods either require complicated hardware implementations or slow z-stack image acquisition. In this work, we develop a new approach to achieve fast autofocusing using the neuromorphic event sensing technology, which detects the sparse brightness changes asynchronously and responds quickly to the specimen movement. A simple yet efficient autofocusing system is tailored for fast acquisition and processing of the non-redundant event data, allowing for autofocusing in only tens of milliseconds, which is thousands of times faster than current technologies. Experimental results show a substantial performance improvement and capability for biopsy specimen inspections. To the best of our knowledge, this is the first reported work on a neuromorphic methodology for autofocusing microscopy.

TMbed: transmembrane proteins predicted through language model embeddings

BackgroundDespite the immense importance of transmembrane proteins (TMP) for molecular biology and medicine, experimental 3D structures for TMPs remain about 4–5 times underrepresented compared to non-TMPs. Today’s top methods such as AlphaFold2 accurately predict 3D structures for many TMPs, but annotating transmembrane regions remains a limiting step for proteome-wide predictions.ResultsHere, we present TMbed, a novel method inputting embeddings from protein Language Models (pLMs, here ProtT5), to predict for each residue one of four classes: transmembrane helix (TMH), transmembrane strand (TMB), signal peptide, or other. TMbed completes predictions for entire proteomes within hours on a single consumer-grade desktop machine at performance levels similar or better than methods, which are using evolutionary information from multiple sequence alignments (MSAs) of protein families. On the per-protein level, TMbed correctly identified 94 ± 8% of the beta barrel TMPs (53 of 57) and 98 ± 1% of the alpha helical TMPs (557 of 571) in a non-redundant data set, at false positive rates well below 1% (erred on 30 of 5654 non-membrane proteins). On the per-segment level, TMbed correctly placed, on average, 9 of 10 transmembrane segments within five residues of the experimental observation. Our method can handle sequences of up to 4200 residues on standard graphics cards used in desktop PCs (e.g., NVIDIA GeForce RTX 3060).ConclusionsBased on embeddings from pLMs and two novel filters (Gaussian and Viterbi), TMbed predicts alpha helical and beta barrel TMPs at least as accurately as any other method but at lower false positive rates. Given the few false positives and its outstanding speed, TMbed might be ideal to sieve through millions of 3D structures soon to be predicted, e.g., by AlphaFold2.

Distributed JPEG Compression and Decompression for Big Image Data Using Map-Reduce Paradigm

Digital data is primarily created and delivered in the form of images and videos in today’s world. Storing and transmitting such a large number of images necessitates a lot of computer resources, such as storage and bandwidth. So, rather than keeping the image data as is, the data could be compressed and then stored, which saves a lot of space. Image compression is the process of removing as much redundant data from an image as feasible while retaining only the non-redundant data. In this paper, the traditional JPEG compression technique is executed in the distributed environment with map-reduce paradigm on big image data. This technique is carried out in serial as well as in parallel fashion with different number of workers in order to show the time comparisons between these setups with the self-created large image dataset. In this, more than one Lakh (121,856) images are compressed and decompressed and the execution times are compared with three different setups: single system, Map-Reduce (MR) with 2 workers and MR with 4 workers. Compression on more than one Million (1,096,704) images using single system and MR with 4 workers is also done. To evaluate the efficiency of JPEG technique, two performance measures such as Compression Ratio (CR) and Peak Signal to Noise Ratio (PSNR) are used.

Optimal Coverage in Wireless Sensor Network using Augmented Nature-Inspired Algorithm

One of the difficult problems that must be carefully considered before any network configuration is getting the best possible network coverage. The amount of redundant information that is sensed is decreased due to optimal network coverage, which also reduces the restricted energy consumption of battery-powered sensors. WSN sensors can sense, receive, and send data concurrently. Along with the energy limitation, accurate sensors and non-redundant data are a crucial challenge for WSNs. To maximize the ideal coverage and reduce the waste of the constrained sensor battery lifespan, all these actions must be accomplished. Augmented Nature-inspired algorithm is showing promise as a solution to the crucial problems in “Wireless Sensor Networks” (WSNs), particularly those related to the reduced sensor lifetime. For “Wireless Sensor Networks” (WSNs) to provide the best coverage, we focus on algorithms that are inspired by Augmented Nature in this research. In wireless sensor networks, the cluster head is chosen using the Diversity-Driven Multi-Parent Evolutionary Algorithm. For Data encryption Improved Identity Based Encryption (IIBE) is used. For centralized optimization and reducing coverage gaps in WSNs Time variant Particle Swarm Optimization (PSO) is used. The suggested model's metrics are examined and compared to various traditional algorithms. This model solves the reduced sensor lifetime and redundant information in Wireless Sensor Networks (WSNs) as well as will give real and effective optimum coverage to the Wireless Sensor Networks (WSNs).

Universal orbit design for metal artifact elimination

Objective. Metal artifacts are a persistent problem in CT and cone-beam CT. In this work, we propose to reduce or even eliminate metal artifacts by providing better sampling of data using non-circular orbits. Approach. We treat any measurements intersecting metal as missing data, and aim to design a universal orbit that can generally accommodate arbitrary metal shapes and locations. We adapted a local sampling completeness metric based on Tuy’s condition to quantify the extent of sampling in the presence of metal. A maxi-min objective over all possible metal locations was used for orbit design. A simple class of sinusoidal orbits was evaluated as a function of frequencies, maximum tilt angles, and orbital extents. Experimental implementation of these orbits were performed on an imaging bench and evaluated on two phantoms, one containing metal balls and the other containing a pedicle screw assembly for spine fixation. Metal artifact reduction (MAR) performance was compared amongst three approaches: non-circular orbits only, algorithmic correction only, and a combined approach. Main results. Theoretical evaluations of the objective favor sinusoidal orbits with large tilt angles and large orbital extents. Furthermore, orbits that leverage redundant azimuthal angles to sample non-redundant data have better performance, e.g. even or non-integer frequency sinusoids for a 360° acquisition. Experimental data support the trends observed in theoretical evaluations. Reconstructions using even or non-integer frequency orbits present less streaking artifacts and background details with finer resolution, even when multiple metal objects are present and even in the absence of MAR algorithms. The combined approach of non-circular orbits and MAR algorithm yields the best performance. The observed trend in image quality is supported by quantitative measures of sampling and severity of streaking artifact. Significance. This work demonstrates that sinusoidal orbits are generally robust against metal artifacts and can provide an avenue for improved image quality in interventional imaging.

Information system of the university editorial-publishing complex

This article presents the results of designing the structure of the data models necessary for the proper functioning of a database supporting the editorial, proofreading, printing, and publishing activities of an educational organization. The goal of the study is to create a coherent, objective, holistic, and non-redundant data set appropriate to the subject area to support the process of digital transformation of the publishing and publication sphere within an educational organization. The resulting data models reflect the logical structure of databases independent of the choice of a particular database management system. As part of the design, JavaScript Object Notation models and relational data models are developed to create a relational database. The results allow identifying and estimating the resources needed to implement each process, determining all participants affecting the implementation of the relevant processes, and defining the access to resources needed by certain participants in the process.

Proteome-pI 2.0: proteome isoelectric point database update.

Proteome-pI 2.0 is an update of an online database containing predicted isoelectric points and pKa dissociation constants of proteins and peptides. The isoelectric point—the pH at which a particular molecule carries no net electrical charge—is an important parameter for many analytical biochemistry and proteomics techniques. Additionally, it can be obtained directly from the pKa values of individual charged residues of the protein. The Proteome-pI 2.0 database includes data for over 61 million protein sequences from 20 115 proteomes (three to four times more than the previous release). The isoelectric point for proteins is predicted by 21 methods, whereas pKa values are inferred by one method. To facilitate bottom-up proteomics analysis, individual proteomes were digested in silico with the five most commonly used proteases (trypsin, chymotrypsin, trypsin + LysC, LysN, ArgC), and the peptides’ isoelectric point and molecular weights were calculated. The database enables the retrieval of virtual 2D-PAGE plots and customized fractions of a proteome based on the isoelectric point and molecular weight. In addition, isoelectric points for proteins in NCBI non-redundant (nr), UniProt, SwissProt, and Protein Data Bank are available in both CSV and FASTA formats. The database can be accessed at http://isoelectricpointdb2.org.

Distributed compression and decompression for big image data: LZW and Huffman coding

In today’s era, digital data are being created and transmitted primarily in the form of images and videos. Storing such a huge number of images and transmitting them requires a lot of computer resources such as storage and bandwidth. So, instead of storing the image data as is, if we compress and store it, it saves a lot of resources. Image compression is the act of removing the maximum possible redundant data from an image and maintaining only the non-redundant data. To compress and decompress such big image data, a distributed environment with a map-reduce paradigm using Hadoop distributed file system and Apache Spark is used. In addition to these, Microsoft Azure cloud environment with infrastructure as a service is also used. Various setups such as a single system, 1 + 4 node cluster, 1 + 15 node cluster, and 1 + 18 node cluster cloud infrastructure are used to show the time comparisons among these setups with the self-created large image dataset. On these four self-made clusters, more than 100 million (109,670,400) images are compressed and decompressed; the execution times are compared with two of the traditional image compression methods: Lempel-Ziv-Welch (LZW) and Huffman coding. Both the LZW and Huffman coding are lossless image compression techniques. LZW removes both spatial and coding redundancies and whereas the Huffman coding removes only coding redundancy. These two compression techniques: LZW and Huffman are just placeholders, these can be replaced with any other compression technique for large image data. In our work, we have used compression ratio, average root mean square error (ARMSE), and average peak signal to noise ratios to validate that the compression and decompression process for each technique is exactly the same irrespective of the number of systems used, distributed or not.

Alignment-Free Antimicrobial Peptide Predictors: Improving Performance by a Thorough Analysis of the Largest Available Data Set.

In the last two decades, a large number of machine-learning-based predictors for the activities of antimicrobial peptides (AMPs) have been proposed. These predictors differ from one another in the learning method and in the training and testing data sets used. Unfortunately, the training data sets present several drawbacks, such as a low representativeness regarding the experimentally validated AMP space, and duplicated peptide sequences between negative and positive data sets. These limitations give a low confidence to most of the approaches to be used in prospective studies. To address these weaknesses, we propose novel modeling and assessing data sets from the largest experimentally validated nonredundant peptide data set reported to date. From these novel data sets, alignment-free quantitative sequence-activity models (AF-QSAMs) based on Random Forest are created to identify general AMPs and their antibacterial, antifungal, antiparasitic, and antiviral functional types. An applicability domain analysis is carried out to determine the reliability of the predictions obtained, which, to the best of our knowledge, is performed for the first time for AMP recognition. A benchmarking is undertaken between the models proposed and several models from the literature that are freely available in 13 programs (ClassAMP, iAMP-2L, ADAM, MLAMP, AMPScanner v2.0, AntiFP, AMPfun, PEPred-suite, AxPEP, CAMPR3, iAMPpred, APIN, and Meta-iAVP). The models proposed are those with the best performance in all of the endpoints modeled, while most of the methods from the literature have weak-to-random predictive agreements. The models proposed are also assessed through Y-scrambling and repeated k-fold cross-validation tests, demonstrating that the outcomes obtained by them are not given by chance. Three chemometric analyses also confirmed the relevance of the peptides descriptors used in the modeling. Therefore, it can be concluded that the models built by fixing the drawbacks existing in the literature contribute to identifying antibacterial, antifungal, antiparasitic, and antiviral peptides with high effectivity and reliability. Models are freely available via the AMPDiscover tool at https://biocom-ampdiscover.cicese.mx/.