Probability Of Correct Prediction Research Articles

Prediction of peptide hormones using an ensemble of machine learning and similarity-based methods.

Peptide hormones serve as genome-encoded signal transduction molecules that play essential roles in multicellular organisms, and their dysregulation can lead to various health problems. In this study, we propose a method for predicting hormonal peptides with high accuracy. The dataset used for training, testing, and evaluating our models consisted of 1174 hormonal and 1174 non-hormonal peptide sequences. Initially, we developed similarity-based methods utilizing BLAST and MERCI software. Although these similarity-based methods provided a high probability of correct prediction, they had limitations, such as no hits or prediction of limited sequences. To overcome these limitations, we further developed machine and deep learning-based models. Our logistic regression-based model achieved a maximum AUROC of 0.93 with an accuracy of 86% on an independent/validation dataset. To harness the power of similarity-based and machine learning-based models, we developed an ensemble method that achieved an AUROC of 0.96 with an accuracy of 89.79% and a Matthews correlation coefficient (MCC) of 0.8 on the validation set. To facilitate researchers in predicting and designing hormone peptides, we developed a web-based server called HOPPred. This server offers a unique feature that allows the identification of hormone-associated motifs within hormone peptides. The server can be accessed at: https://webs.iiitd.edu.in/raghava/hoppred/.

Outcome Supervised Deep Learning Model on Pathological Whole Slide Images for Survival Prediction of Immunotherapy in Non-Small Cell Lung Cancer Patients: A Multicenter Study

Although PD-(L)1 inhibitors were marked by durable efficacy in non-small cell lung cancer patients (NSCLC), about 60% of patients still suffer from recurrence and metastasis after PD-(L)1 inhibitors treatment. And there were no robust biomarkers of the response of PD-(L)1 inhibitors. Whole slide images (WSIs) of H&E-stained specimens have been found to characterize the tumor microenvironment, and might be the potential prognostic predictors of NSCLC patients. To accurately predict the response to PD-(L)1 inhibitors, we presented the deep learning model based on WSI of H&E-stained specimens of NSCLC patients. Two independent cohorts of NSCLC patients receiving PD-(L)1 inhibitors from two hospitals were enrolled for model training and testing respectively. The WSI images of H&E-stained histological specimens were obtained from these patients, and patched into 1024×1024 pixels. The labels of patched images were determined due to their progression free survival (PFS) with the interval of 4 months. The patch-level model was firstly trained based on Vit to identify the predictive patches in training cohort, and patch-level probability distribution was performed. Then we trained patient-level survival model-based Vit-RNN framework, and tested it in external validation cohort. A total of 291 WSI images of H&E-stained histological specimens from 198 NSCLC patients in primary cohort and 62 WSI images from 30 NSCLC patients in testing cohort were included for model training and external validation. All patients were divided into 4 groups due to their PFS after PD-(L)1 inhibitors. There were 246,318 patches from 291 images in primary cohort after image pre-processing, and all images were randomly divided into train cohort and validation cohort with the proportion of 7:3. The patch-level Vit model with the highest accuracy was saved and the predictive patches were selected after 50 epochs training. All patches were ranked by the probability of correct prediction, and the first 50 top-ranked patches from each WSI image are sequentially passed to the patient-level Vit-RNN model. The Vit-RNN survival achieved an accuracy of 88.6% in the validation cohort, and an accuracy of 81% in the testing cohort. The multivariate cox analysis also indicated the Vit-RNN survival model remained a statistically independent predictor of survival from PD-(L)1 inhibitors (P = 0.0085). The outcome supervised Vit-RNN survival model based on pathological WSIs could be used to predict the efficacy the PD-(L)1 inhibitors in NSCLC patients, laying the foundation for the deployment of computational pathomics in clinical practice of immunotherapy.

Big Data analytics for improved prediction of ligand binding and conformational selection.

This research introduces new machine learning and deep learning approaches, collectively referred to as Big Data analytics techniques that are unique to address the protein conformational selection mechanism for protein:ligands complexes. The novel Big Data analytics techniques presented in this work enables efficient data processing of a large number of protein:ligand complexes, and provides better identification of specific protein properties that are responsible for a high probability of correct prediction of protein:ligand binding. The GPCR proteins ADORA2A (Adenosine A2a Receptor), ADRB2 (Adrenoceptor Beta 2), OPRD1 (Opioid receptor Delta 1) and OPRK1 (Opioid Receptor Kappa 1) are examined in this study using Big Data analytics techniques, which can efficiently process a huge ensemble of protein conformations, and significantly enhance the prediction of binding protein conformation (i.e., the protein conformations that will be selected by the ligands for binding) about 10-38 times better than its random selection counterpart for protein conformation selection. In addition to providing a Big Data approach to the conformational selection mechanism, this also opens the door to the systematic identification of such "binding conformations" for proteins. The physico-chemical features that are useful in predicting the "binding conformations" are largely, but not entirely, shared among the test proteins, indicating that the biophysical properties that drive the conformation selection mechanism may, to an extent, be protein-specific for the protein properties used in this work.

Assessment of the relationship between regional wall motion abnormality score revealed by parametric imaging and the extent of LGE with CMR

The aim of this study was to assess the relationship between left ventricular (LV) regional myocardial wall motion abnormality (WMA), revealed by visual interpretation of cardiac magnetic resonance (CMR) cine images together with the computed wall motion parametric image, and the transmural scar extent, as assessed by Late gadolinium Enhancement (LGE), in 40 patients. Each cine CMR short-axis loop was processed to compute a parametric image where each pixel represents the amplitude of the Hilbert transform of videointensity over time. Two expert radiologists blindly interpreted the cine CMR images in combination with the corresponding parametric image to assign a WMA score for each of the 16 myocardial sectors in which the LV myocardium was subdivided. Such score was compared per sector to the level of transmural scar extent obtained by LGE images.A total of 592 myocardial segments were analyzed. A significant decrease in regional wall motion was observed in sectors with LGE transmural hyperenhancement > 75% of tissue, as well as a correlation between parametric image amplitude and peak radial and circumferential strain, computed by feature tracking. The results showed a reduction in prediction error Lambda of WMA from LGE of 65%, and of LGE from WMA of 63%. In particular, the estimated probability of correct prediction of WMA from LGE was 76%, while that of LGE from WMA was 75%.The interpretation of myocardial viability by LGE images combined with the WMA information, derived from cine CMR and parametric images, could improve the clinical decision making process.

ToxinPred2: an improved method for predicting toxicity of proteins.

Proteins/peptides have shown to be promising therapeutic agents for a variety of diseases. However, toxicity is one of the obstacles in protein/peptide-based therapy. The current study describes a web-based tool, ToxinPred2, developed for predicting the toxicity of proteins. This is an update of ToxinPred developed mainly for predicting toxicity of peptides and small proteins. The method has been trained, tested and evaluated on three datasets curated from the recent release of the SwissProt. To provide unbiased evaluation, we performed internal validation on 80% of the data and external validation on the remaining 20% of data. We have implemented the following techniques for predicting protein toxicity; (i) Basic Local Alignment Search Tool-based similarity, (ii) Motif-EmeRging and with Classes-Identification-based motif search and (iii) Prediction models. Similarity and motif-based techniques achieved a high probability of correct prediction with poor sensitivity/coverage, whereas models based on machine-learning techniques achieved balance sensitivity and specificity with reasonably high accuracy. Finally, we developed a hybrid method that combined all three approaches and achieved a maximum area under receiver operating characteristic curve around 0.99 with Matthews correlation coefficient 0.91 on the validation dataset. In addition, we developed models on alternate and realistic datasets. The best machine learning models have been implemented in the web server named 'ToxinPred2', which is available at https://webs.iiitd.edu.in/raghava/toxinpred2/ and a standalone version at https://github.com/raghavagps/toxinpred2. This is a general method developed for predicting the toxicity of proteins regardless of their source of origin.

Predictability of Internet of Things Traffic at the Medium Access Control Layer Against Information-Theoretic Bounds

Most of the existing Medium Access Control (MAC) layer protocols for the Internet of Things (IoT) model the traffic generated by each IoT device via random arrivals such as those in a Poisson process. Under this model, since it is implied that IoT device traffic cannot be predicted, only reactive MAC-layer protocols in which the network responds to the current traffic are viable. In contrast, recent work has demonstrated that the traffic generated by an individual IoT device can be predictable, thus enabling predictive network protocols at the MAC layer. In this paper, we investigate information-theoretic bounds on the predictability of IoT traffic of individual devices. To this end, first, we compare the performance achieved by the following state-of-the-art forecasters on individual IoT device traffic: Logistic Regression, Multi-Layer Perceptron (MLP), 1-Dimensional Convolutional Neural Network (1D CNN), and Long Short Term Memory (LSTM) as well as MLP under feature selection based on Analysis of Variance (ANOVA) and Auto-Correlation Function (ACF). Second, we quantify the gap between the performance of these forecasters against information-theoretic bounds as follows: For IoT devices that generate a fixed number of bits at each generation instance, we measure the gap between the forecasting accuracy and the information-theoretic bound established by Fano’s inequality on the probability of correct prediction. Our empirical results show that existing forecasting schemes perform close to the information-theoretic bound in this case. For IoT devices that generate a variable number of bits, we measure the gap between the Mean Square Error (MSE) and the estimation-theoretic counterpart to Fano’s inequality. Our empirical results show that the performance of existing forecasting schemes is far from the information-theoretic bound in this case. This work thus motivates the machine learning community to develop forecasting schemes that approach information-theoretic bounds. Furthermore, this work is expected to impact the development of predictive MAC-layer protocols that exploit these bounds.

Iterative Named Entity Recognition with Conditional Random Fields

Named entity recognition (NER) constitutes an important step in the processing of unstructured text content for the extraction of information as well as for the computer-supported analysis of large amounts of digital data via machine learning methods. However, NER often relies on domain-specific knowledge, being conducted manually in a time- and human-resource-intensive process. These can be reduced with statistical models performing NER automatically. The current work investigates whether Conditional Random Fields (CRF) can be efficiently trained for NER in German texts, by means of an iterative procedure combining self-learning with a manual annotation–active learning–component. The training dataset increases continuously with the iterative procedure. Whilst self-learning did not markedly improve the performance of the CRF for NER, the manual annotation of sentences with the lowest probability of correct prediction clearly improved the model F1-score and simultaneously reduced the amount of manual annotation required to train the model. A model with an F1-score of 0.885 was able to be trained in 11.4 h.

Transmission Line Galloping Prediction Based on GA-BP-SVM Combined Method

The paper proposes a method to construct a model for transmission line galloping prediction using machine learning algorithms to address the transmission line galloping problem, which can result in transmission line loss and pose a greater risk to the safety of electricity in society. First, to reduce sensor noise interference, a unilateral sliding time window is used for micro-meteorological data correction, and then gray correlation analysis method and the specific gravity method are used to obtain the influence weights of micro-meteorological elements. The galloping prediction models are constructed using six algorithms, with the GA-BP algorithm model and the SVM algorithm model having better prediction effects based on performance metrics. The GA-BP-SVM combined model is constructed on this basis, and all of its performance metrics are optimal. This model's prediction accuracy in both galloping and no galloping states reaches 95.5%; the probability of correct prediction when predicted as galloping reaches 95.1%; the probability that actual galloping can be predicted reaches 92.5%. The F1-score of the combined model reaches 0.938, which indicates that it has the best prediction effect. The prediction method described in the paper is accurate and practical, and operation and maintenance personnel can flexibly develop inspection strategies and anti-galloping measures based on the prediction results to ensure the safe and stable operation of transmission lines.

PAC learning of arbiter PUFs

The general concept of physically unclonable functions (PUFs) has been nowadays widely accepted and adopted to meet the requirements of secure identification and key generation/storage for cryptographic ciphers. However, shattered by different attacks, e.g., modeling attacks, it has been proved that the promised security features of arbiter PUFs, including unclonability and unpredictability, are not supported unconditionally. However, so far the success of existing modeling attacks relies on pure trial and error estimates. This means that neither the probability of obtaining a useful model (confidence), nor the sufficient number of CRPs, nor the probability of correct prediction (accuracy) is guaranteed. To address these issues, this work presents a probably approximately correct (PAC) learning algorithm. Based on a crucial discretization process, we are able to define a Deterministic finite automaton (of polynomial size), which exactly accepts the regular language corresponding to the challenges mapped by the given PUF to one responses.

Improved exon prediction with transforms by de-noising period-3 measure

Gene finding techniques in eukaryotic cells can be divided into two categories, viz. - model-dependent and model-independent. In model-independent category, transforms are commonly used to identify exons or genes present in DNA sequences. In this work, a Post-Processing Algorithm (PPA) for enhancing gene prediction features of transforms is developed. PPA compares the N/3 spectral components of DNA signal with the corresponding spectrum of period-3 suppressed DNA signal. In the N/3 spectrum of DNA sequences, the bases for which the difference between these two spectrums is within a predefined threshold level are marked as non-coding (introns) regions. In such regions the signal values are replaced by the difference signal of the two spectrums. This substitution suppresses the noise in the intronic regions of the N/3 spectrum; while the coding region (exonic) signals are not affected, resulting in de-noised period-3 measures. PPA has been applied to process the period-3 coefficients of Discrete Fourier Transform (DFT), Paired Spectral Content (PSC), and Modified Gabor Wavelet Transform (MGWT) methods to de-noise their period-3 measures. Performance of the algorithm has been evaluated on HMR195, Burset/Guigo570, and Asp67 datasets using Receiver Operating Characteristic (ROC) and specificity versus sensitivity curves. The PPA, while preserving the model-independent characteristic of transform based methods, improves the probability of correct prediction of the exonic regions.

Accuracy Assessment of Sequential Indicator Simulation in Three-dimensional Prediction of Soil Texture

Equiprobable realizations of soil textural maps can be drawn using Sequential Indicator Simulation (SIS), which reflects the probability of occurrence of each texture and is constrained by the observed textures at the observation sites. However, the SIS is not an error-free technique, and the accuracy of these maps should be checked before they are used as basic information for precision agricultural- and environmental-related studies. This article assesses the accuracy of using SIS in the three-dimensional prediction of soil texture. A soil data set (139 profiles) with five types of textures distributed in a 15-km region was first collected and then randomly sub-divided into a training set (85 profiles) and a validation set (54 profiles). Second, 100 realizations were obtained by SIS using the training set. Finally, the prediction capacity was assessed using independent validation set and probability of correct prediction as criterion. Results show that 43.59% of total observations can be correctly predicted while the accuracy varies among textures and depths. The dominant textures in the data set have higher accuracy (>42.49%), while the textures with less proportion (<28.86%) were poorly predicted. The SIS performed better for the near-surface depth (0–0.5 m) than deeper depths (0.5–2.0 m). Therefore, further improvement in simulation of soil texture is necessary as correct predictions of these minor textures and deeper depth textures were very low.

Improving the detection of transmembrane β-barrel chains with N-to-1 extreme learning machines

Transmembrane β-barrels (TMBBs) are extremely important proteins that play key roles in several cell functions. They cross the lipid bilayer with β-barrel structures. TMBBs are presently found in the outer membranes of Gram-negative bacteria and of mitochondria and chloroplasts. Loop exposure outside the bacterial cell membranes makes TMBBs important targets for vaccine or drug therapies. In genomes, they are not highly represented and are difficult to identify with experimental approaches. Several computational methods have been developed to discriminate TMBBs from other types of proteins. However, the best performing approaches have a high fraction of false positive predictions. In this article, we introduce a new machine learning approach for TMBB detection based on N-to-1 Extreme Learning Machines that significantly outperforms previous methods achieving a Matthews correlation coefficient of 0.82, a probability of correct prediction of 0.92 and a sensitivity of 0.73.

Prediction of Long-Term Canker Disease Damage from the Responses of Juvenile Poplar Clones to Inoculation with Septoria musiva.

Stem cankers caused by Septoria musiva severely limit production of susceptible hybrid poplars in eastern North America. A field experiment was conducted to determine whether short-term responses of poplar stems to inoculation with S. musiva were predictive of long-term canker disease damage. Stems of 27 poplar clones were inoculated during their first season of growth by removing the fourth or fifth fully expanded leaf and placing an agar plug colonized by an aggressive isolate of S. musiva over the resulting wound. Four months after inoculation, incidence of cankers, canker length, and percent of stem circumference affected (girdle) were recorded. Clones varied greatly in canker incidence (12 to 98%), mean canker length (10 to 53 mm), and mean girdle (14 to 94%). Logistic regression analysis was used to compare these inoculation responses with canker disease damage categories assigned on the basis of information from longer-term field studies. Incidence, canker length, and girdle data were all informative, but girdle data from this field experiment correctly predicted assigned canker disease damage categories most frequently (24 of 27 clones). Responses of 15 of these clones also were evaluated in a similar greenhouse experiment. Although responses usually were predictive of long-term damage categories, the probability of correct prediction was lower than that obtained in the field experiment for most clones. These results demonstrate the feasibility and potential benefit of screening juvenile poplar clones for responses to inoculation with S. musiva before undertaking field trials for productivity and subsequent release to growers.

Predicting alpha-helix and beta-strand segments of globular proteins.

All current methods of protein secondary structure prediction are based on evaluation of a single residue state. Although the accuracy of the best of them is approximately 60-70%, for reliable prediction of tertiary structure it is more useful to predict an approximate location of alpha-helix and beta-strand segments, especially prolonged ones. We have developed a simple method for protein secondary structure prediction which is oriented on the location of secondary structure segments. The method uses linear discriminant analysis to assign segments of a given amino acid sequence a particular type of secondary structure, by taking into account the amino acid composition of internal parts of segments as well as their terminal and adjacent regions. Four linear discriminant functions were constructed for recognition of short and long alpha-helix and beta-strand segments respectively. These functions combine three characteristics: hydrophobic moment, segment singlet, and pair preferences to an alpha-helix or beta-strand. The last two characteristics are calculated by summing the preference parameters of single residues and pairs of residues located in a segment and its adjacent regions. The final program SSP predicts all possible potential alpha-helices and beta-strands and resolves some possible overlap between them. Overall three-state (alpha, beta, c) prediction gives approximately 65.1% correctly predicted residues on 126 non-homologous proteins using the jackknife test procedure. Analysis of the prediction results shows a high prediction accuracy of long secondary structure segments (approximately 89% of alpha-helices of length > 8 and approximately 71% of beta-strands of length > 6 are correctly located with probability of correct prediction 0.82 and 0.78 respectively.(ABSTRACT TRUNCATED AT 250 WORDS)

Systematic study and periodic dependence of equilibrium distribution coefficients of admixtures on the atomic number of admixtures and their importance for crystallization

A systematic study is presented of equilibrium and liquidus curves of available binary systems A-M, where A are basic elements and M are admixtures, for 50 basic elements. The values of equilibrium distribution coefficients of admixtures koM A in dilute solution in the range near the pure basic metals were calculated. The correlation curves of koM A vs. the atomic number were constructed. These correlations enable us to predict the unknown koM A. The different behaviour of admixtures with complete and incomplete d electron levels was shown. Two independent parabolic approximations describing the periodic dependence of koM A on the atomic number of admixtures for the basic elements Si, Ge, Ti, Zr and Hf were investigated. There is a relatively good agreement between measured and calculated values of koM A. The probability of correct prediction of the unknown koM A for some admixtures has increased accordingly.

A statistical approach to learning and generalization in layered neural networks

A general statistical description of the problem of learning from examples is presented. Learning in layered networks is posed as a search in the network parameter space for a network that minimizes an additive error function of a statistically independent examples. By imposing the equivalence of the minimum error and the maximum likelihood criteria for training the network, the Gibbs distribution on the ensemble of networks with a fixed architecture is derived. The probability of correct prediction of a novel example can be expressed using the ensemble, serving as a measure to the network's generalization ability. The entropy of the prediction distribution is shown to be a consistent measure of the network's performance. The proposed formalism is applied to the problems of selecting an optimal architecture and the prediction of learning curves.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

Probability processing and diagnostic search: 20 alternatives, 500 trials

Ss’ capabilities for processing probability information in performing a diagnostic search task similar to troubleshooting is explored. Six Ss searched (queried) among 20 components in a series circuit to isolate a single component failure per trial. Failures were generated according to 10 alternative probability-of-component-failure distributions (D), each with N = 500. Prior to performing their search on a trial, Ss predicted the failure for that trial. Of primary interest was the influence of varying D on predictions and search performance. Results indicate that: (1) distributions of predictions tend to match actual D in shape and location but exhibit certain characteristics; (2) probability of correct prediction is low and is influenced by D; (3) search strategies are suboptimal but “logical” and to some extent predictable; and (4) in some instances, search strategy appears to be related to predictions.