Discovery Probability Research Articles

How to explore chemical space using algorithms and automation

Although extending the reactivity of a given class of molecules is relatively straightforward, the discovery of genuinely new reactivity and the molecules that result is a wholly more challenging problem. If new reactions can be considered unpredictable using current chemical knowledge, then we suggest that they are not merely new but also novel. Such a classification, however, requires an expert judge to have access to all current chemical knowledge or risks a lack of information being interpreted as unpredictability. Here, we describe how searching chemical space using automation and algorithms improves the probability of discovery. The former enables routine chemical tasks to be performed more quickly and consistently, while the latter uses algorithms to facilitate the searching of chemical knowledge databases. Experimental systems can also be developed to discover novel molecules, reactions and mechanisms by augmenting the intuition of the human expert. In order to find new chemical laws, we must seek to question current assumptions and biases. Accomplishing that involves using two areas of algorithmic approaches: algorithms to perform searches, and more general machine learning and statistical modelling algorithms to predict the chemistry under investigation. We propose that such a chemical intelligence approach is already being used and that, in the not-too-distant future, the automated chemical reactor systems controlled by these algorithms and monitored by a sensor array will be capable of navigating and searching chemical space more quickly, efficiently and, importantly, without bias. This approach promises to yield not only new molecules but also unpredictable and thus novel reactivity. Automation can help in performing routine tasks quickly and consistently. Algorithms facilitate the searching of current knowledge. Combining the two could lead to a chemically intelligent approach to the discovery of not only new molecules but also novel and unpredictable reactivity.

A BP Neural Network Prediction Model Based on Dynamic Cuckoo Search Optimization Algorithm for Industrial Equipment Fault Prediction

The fault prediction problem for modern industrial equipment is a hot topic in current research. So, this paper first proposes a dynamic cuckoo search algorithm. The algorithm improves the step size and discovery probability. Then, it introduces the change trend of fitness function value into the step size update formula to balance the search speed and accuracy. At the same time, the algorithm initial global search step is larger, while the step size of the local search is smaller in the latter part of the algorithm. In the process of discovering the global optimal solution, the probability of preserving the offspring with good fitness is increased, and the uncertainty of preference random walk is improved. As the search progresses, the probability of discovery is reduced, which makes it easy to produce new individuals in the later stage of evolution. Based on this, a BP neural network prediction model based on dynamic cuckoo search algorithm optimization is established. And the experimental results show that the proposed prediction model has faster convergence and higher accuracy.

Bayesian statistical methods in genetic association studies: Empirical examination of statistically non-significant Genome Wide Association Study (GWAS) meta-analyses in cancers: A systematic review

A Bayesian statistical method was developed to assess the noteworthiness of a single nucleotide polymorphism (SNP)-phenotype association that shows statistical significance in various observational studies, but it has seldom been applied to GWAS meta-analyses in cancers. Data (i.e. allelic frequency, odds ratio, 95% confidence interval, etc.) on various SNP-cancer associations were extracted from meta-analysis of GWAS and the National Human Genome Research Institute (NHGRI) Catalog of Published GWAS and were used to compute the false positive report probability (FPRP) and Bayesian false discovery probability (BFDP) to evaluate the noteworthiness of SNP-cancer associations. Independent paired t-tests showed a direct relationship between SNP-cancer P-values and both FPRP and BFDP estimates. However, a discrepancy in the number of noteworthy associations between P-value comparison and either FPRP or BFDP was found using data extracted from meta-analyses of GWAS and the GWAS Catalog. Most P-values of associations with nonsignificant P-values but with noteworthy FPRP and BFDP estimates were within the range of 10−6 to 5 × 10−8. A poorly selected genome-wide significance threshold and inclusion of a nonsignificant SNP-phenotype association into the noteworthy test can, with either noteworthy FPRP or BFDP computation, give a false impression of noteworthiness for a nonsignificant association.

Meta-Analysis of Genome-Wide Association Studies (GWAS) of Late Toxicity in 3,874 Men Treated with Radiation for Prostate Cancer

Late toxicity following prostate cancer radiotherapy negatively impacts quality of life and can be dose-limiting. Radiosensitivity is a genetic trait, and identification of single nucleotide polymorphisms (SNPs) associated with toxicity could uncover novel radiobiology pathways and contribute to precision cancer care. The largest study to date meta-analyzed data from 1,564 men with prostate cancer and identified two risk SNPs. This expanded study added new cohorts to greatly increase the sample to 3,874 men with the aim of identifying new toxicity risk SNPs and validating previously identified SNPs. A GWAS meta-analysis was performed with six independent cohorts of men treated with radiotherapy for prostate cancer: RAPPER (n=2,010), RADIOGEN (n=658), GenePARE (n=495), UGhent (n=311), CCI-BT (n=252), and CCI-EBRT (n=148). Four toxicities were assessed prospectively from 6 months after radiotherapy and harmonized to a common scale: increased urinary frequency, decreased urine stream, hematuria, and rectal bleeding. Germline DNA was genotyped using genome-wide arrays. Imputation generated a uniform set of ∼7 million SNPs (frequency in Europeans >4%) across cohorts. Cox proportional hazards regression tested association between each SNP and time to first occurrence of grade 2 or worse toxicity within each cohort. Individual cohort results were combined using inverse variance weighted fixed-effects meta-analysis. Novel SNP-toxicity associations were considered significant if they had meta-p-value <5x10-8. Bayesian False Discovery Probability (BFDP) provided an additional measure of confidence. Five new risk loci for late toxicity were identified in 3,874 prostate cancer patients: rs17055178 (HR 1.98; 95%CI 1.60 to 2.44; p=2.4x10-10; BFDP <1%) and rs9644474 (HR 0.47; 95%CI 0.36 to 0.62; p=4.3x10-8; BFDP 10%) associated with rectal bleeding; rs10969913 (HR 3.82; 95%CI 2.50 to 5.83; p=6.2x10-10; BFDP 1.6%) with decreased urine stream; rs11122573 (HR 1.89; 95%CI 1.50 to 2.38; p=2.3x10-8; BFDP 5.7%) and rs139882217 (HR 4.83; 95%CI 2.73 to 8.52; p=2.7x10-8; BFDP 69%) with hematuria. Previously published risk SNPs showed consistent associations in the cohorts not included in prior discovery GWAS: rs1801516 in ATM associated with overall toxicity (ß=0.05; 95%CI 0.01 to 0.09; p=0.016), rs17599026 in KDM3B associated with 2yr prevalence of increased urinary frequency (OR 1.27; 95%CI 0.95 to 1.70; p=0.10), rs7720298 in DNAH5 associated with 2yr prevalence of decreased urine stream (OR 1.26; 95%CI 0.94 to 1.70; p=0.13). A locus in TANC1 showed replication in some, but not all, new cohorts and warrants further investigation. This largest GWAS to date identified novel risk SNPs and replicated previously identified SNPs. Further expansion of cohorts will uncover additional SNPs and allow development of risk models. Such models can be tested in clinical trials aimed at stratification of cancer patients based on toxicity risk profile.

Incremental Science: Papers and Grants, Yes; Discoveries, No

Today, the effectiveness of scientific research is increasingly often determined by the size of grants obtained, the number of publications, and the level of their prestige, as well as the ability to attract public attention, rather than by well-defined problems and stringent methods of their solutions. This trend is disastrous for science. Bibliometric data, useful as a subsidiary method of evaluation, cannot replace expert opinions based on the analysis of the content of the paper. The pressure making scientists publish is leading to poor quality publications. The irreproducibility of research papers, amounting in oncology up to 75%, is now commonplace. This results in an incremental character of science, when the research is reduced to a systematic refinement of the available data. As a consequence, researchers and financing focus on known and common research directions, while there are still many understudied areas avoided by researchers due to the risk of failure and loss of publications. For instance, 1000 of 4000 E. coli genes remain absolutely uninvestigated. Clearly, there are even more unknown areas in complex multicellular organisms, including humans. This has created prerequisites to reduce the probability of large scientific discoveries, which are often made by chance. Science needs better financing and this is aggravated by the fact that the allocated funds are distributed without considering the scientific value of research and the accidental character of discoveries. The author draws attention to the importance of financial support of scientific schools that are the most probable centers for obtaining new scientific knowledge and preparing the ground for scientific discoveries.

Genetic variation and systemic lupus erythematosus: A field synopsis and systematic meta-analysis.

Systemic lupus erythematosus (SLE) is a multi-systemic severe autoimmune disease which results from the irreversible loss of self-tolerance and impaired molecular responses, especially an altered interferon signature. We synthesized all meta-analyses reporting a genetic association of SLE, and further investigated their validity to discover false positive results under Bayesian methods. We executed a PubMed search to extract the respective results regarding gene polymorphisms of SLE, published until June 30th 2017 and selected a single result per genetic variant among duplicates. Among 133 significant genotype comparisons, 45 (34%) were found noteworthy under both false positive report probability (FPRP) and Bayesian false discovery probability (BFDP). From the meta-analysis of genome-wide association studies (GWAS), we could confirm that all significant comparisons were noteworthy under both Bayesian approaches. Both approaches may be advantageous for determining whether the reported associations are genuine, especially for interpreting results from observational studies instead of GWAS whose significance was determined in a more strict manner. When determining results from GWAS with a p-value ranging between 0.05 and 5 × 10-8, other statistical approaches, rather than single standard significance may be beneficial. Taking into account these considerations, a proportion of meta-analyses claimed statistical significance, but these results need to be interpreted with caution.

Neighbourhood‐aware out‐of‐network D2D discovery

Device-to-device communication in 5G networks will require fast and efficient neighbour discovery. This Letter shows that with an initial knowledge of proximity, devices can increase discovery probability. However, in out-of-coverage scenarios, devices do not have information about neighbourhood which decreases the discovery probability. Hence, this Letter proposes an algorithm that enables neighbourhood-awareness in out-of-coverage environments, and increases discovery probability.

Modified cuckoo search algorithm for the optimal placement of actuators problem

This paper proposes a novel modified cuckoo search algorithm (NMCSA) to solve optimal placement of actuators (OPA) for active vibration control. The purpose of OPA is to minimize control spillover effect and maximize the control force applied to the desired modes. To achieve this objective, NMCSA first employs speed factor (SFR) and aggregation factor (AFR) for recording and analyzing the current and history information of nests. Secondly, SFR and AFR are mapped to suitable space by scale conversion factors (SCF). Thus, the NMCSA based on SCF can give adaptively actions on the step size α and discovery probability pa to balance exploration and exploitation. The performance of NMCSA is confirmed by some well-known benchmark functions. Subsequently, the NMCSA is applied to solve OPA and compared with several state-of-the-art algorithms in the literature, the statistical results demonstrate that the proposed algorithm has a higher convergence speed and better search ability.

Connectivity Times in Vehicular Networks

We propose an analytical model that accurately captures the service advertisement and access mechanisms in IEEE 1609.4/IEEE 802.11p multi-channel vehicular networks where road-side units announce non-safety services to passing vehicles on a dedicated channel. For a drive-thru scenario, we calculate various performance measures including the service discovery probability, the mean time till discovery, and the channel utilization.

Mining for Small Translated ORFs.

Peptides encoded by short open reading frames (sORFs) are usually defined as peptides ≤100 aa long. Usually sORFs were ignored by automatic genome annotation programs due to the high probability of false discovery. However, improved computational tools along with a high-throughput RIBO-seq approach identified a myriad of translated sORFs. Their importance becomes evident as we are gaining experimental validation of their diverse cellular functions. This Review examines various computational and experimental approaches of sORFs identification as well as provides the summary of our current knowledge of their functional roles in cells.

When does monitoring hurt? Endogenous information acquisition in a game of persuasion

We study a persuasion game between a decision maker and an expert. Before communication, the expert searches for an evidence about the state of nature. Higher effort leads to a higher probability of discovery. We show that, contrary to everyday experiences, monitoring of effort is always suboptimal if the expert’s bias is large, yet sometimes optimal if the expert’s bias is small.

Novel Device-to-Device Discovery Scheme Based on Random Backoff in LTE-Advanced Networks

Device-to-device (D2D) discovery is a key enabler of D2D communications for the direct exchange of local area traffic between proximity users (UEs) to improve spectral efficiency. The direct D2D discovery relies on the capabilities of the D2D UEs to autonomously indicate their presence to proximity D2D UEs. Despite its potential of reducing energy and signaling burden, the direct D2D discovery has not drawn adequate attention. In this paper, we propose a direct D2D discovery scheme based on the random backoff procedure, where D2D UEs randomly choose a backoff interval and retransmit a beacon. Compared with existing schemes, the performance of the proposed scheme can be significantly enhanced in terms of the discovery probability and the discovery delay. Several useful guidelines for its design are proposed based on our analysis. Finally, numerical results provide valuable insights on the performance tradeoff inherent to the proposed D2D discovery scheme.

An improved cuckoo search algorithm and its application in vibration fault diagnosis for a hydroelectric generating unit

ABSTRACTThis article presents an improved cuckoo search (ICS) algorithm based on the parametric adaptation mechanism and non-uniform mutation. In the ICS algorithm, the biased random walk operator is modified to take advantage of the neighbourhood information of the current solution, and then the control parameters of step size, discovery probability and scaling factor are directly integrated into the optimized problems. Meanwhile, the non-uniform mutation operation is used to adaptively tune the search step of the current optimal solution. To evaluate the feasibility and efficiency of the presented algorithm, three groups of benchmark test functions are employed to perform the validation analysis. Subsequently, to diagnose the vibration fault of a hydroelectric generating unit, a combinational model is built, which combines ICS with a back-propagation neural network (ICSBP). The experimental results indicate that ICS is competitive on the optimization problems, and the ICSBP approach can effectively improve the accuracy of vibration fault diagnosis for a hydroelectric generating unit.

Field Synopsis and Re-analysis of Systematic Meta-analyses of Genetic Association Studies in Multiple Sclerosis: a Bayesian Approach.

To provide an up-to-date summary of multiple sclerosis-susceptible gene variants and assess the noteworthiness in hopes of finding true associations, we investigated the results of 44 meta-analyses on gene variants and multiple sclerosis published through December 2016. Out of 70 statistically significant genotype associations, roughly a fifth (21%) of the comparisons showed noteworthy false-positive rate probability (FPRP) at a statistical power to detect an OR of 1.5 and at a prior probability of 10-6 assumed for a random single nucleotide polymorphism. These associations (IRF8/rs17445836, STAT3/rs744166, HLA/rs4959093, HLA/rs2647046, HLA/rs7382297, HLA/rs17421624, HLA/rs2517646, HLA/rs9261491, HLA/rs2857439, HLA/rs16896944, HLA/rs3132671, HLA/rs2857435, HLA/rs9261471, HLA/rs2523393, HLA-DRB1/rs3135388, RGS1/rs2760524, PTGER4/rs9292777) also showed a noteworthy Bayesian false discovery probability (BFDP) and one additional association (CD24 rs8734/rs52812045) was also noteworthy via BFDP computation. Herein, we have identified several noteworthy biomarkers of multiple sclerosis susceptibility. We hope these data are used to study multiple sclerosis genetics and inform future screening programs.

Discovery probability of next-generation neutrinoless double- β decay experiments

The Bayesian discovery probability of future experiments searching for neutrinoless double-$\beta$ decay is evaluated under the popular assumption that neutrinos are their own antiparticles. A Bayesian global fit is performed to construct a probability distribution for the effective Majorana mass, the observable of interest for these experiments. This probability distribution is then combined with the sensitivity of each experiment derived from a heuristic counting analysis. The discovery probability is found to be higher than previously considered, but strongly depends on whether the neutrino mass ordering is normal or inverted. For the inverted ordering, next-generation experiments are likely to observe a signal already during their first operational stages. Even for the normal ordering, in the absence of neutrino mass mechanisms that drive the lightest state or the effective Majorana mass to zero, the probability of discovering neutrinoless double-$\beta$ decay can reach $\sim$50% or more in the most promising experiments.

Study of the Galactic Center region with the ANTARES neutrino telescope

The measurements of astrophysical neutrinos by the IceCube collaboration contains some indications of a North/South asymmetry which could hint a Galactic contribution. The ANTARES neutrino telescope has a direct view of the Galactic Center region and can provide complementary information on the neutrino flux from this region thanks to its excellent angular resolution both for tracks and showers. The recent model KRAγ, characterized by radially-dependent cosmic-ray transport properties predicts a neutrino flux close to the ANTARES sensitivity. We present here a study on a possible Galactic contribution to the astrophysical neutrino flux using seven years of data from the ANTARES neutrino telescope. The results of this analysis are preliminary, the median upper limit at 90% confidence level is roughly two times larger than the flux predicted by the KRAγ model and the discovery probability at 3σ is 3% of the KRAγ flux.

Modified cuckoo search algorithm and the prediction of flashover voltage of insulators

A novel modified cuckoo search (MCS) algorithm is proposed to enhance the performance of cuckoo search. In this MCS algorithm, an enhancing exploration strategy is adopted to extend the search region. At the same time, the control parameters of step size and discovery probability are adjusted adaptively according to the improvement rate of the solutions toward the optimal value. To evaluate the effectiveness of the presented MCS algorithm, the simulation experiment is executed on 23 benchmark test functions. The results demonstrate that MCS algorithm performs better than the others reported in the literature. Moreover, a combination algorithm, which the MCS combined with BP neural network, is utilized to predict the flashover voltage on polluted insulators. The comparative study results further indicate that MCS is more efficient and effective, and the combination algorithm is suitable for the prediction of flashover voltage on insulators.

Proposal and Evaluation of BLE Discovery Process Based on New Features of Bluetooth 5.0.

The device discovery process is one of the most crucial aspects in real deployments of sensor networks. Recently, several works have analyzed the topic of Bluetooth Low Energy (BLE) device discovery through analytical or simulation models limited to version 4.x. Non-connectable and non-scannable undirected advertising has been shown to be a reliable alternative for discovering a high number of devices in a relatively short time period. However, new features of Bluetooth 5.0 allow us to define a variant on the device discovery process, based on BLE scannable undirected advertising events, which results in higher discovering capacities and also lower power consumption. In order to characterize this new device discovery process, we experimentally model the real device behavior of BLE scannable undirected advertising events. Non-detection packet probability, discovery probability, and discovery latency for a varying number of devices and parameters are compared by simulations and experimental measurements. We demonstrate that our proposal outperforms previous works, diminishing the discovery time and increasing the potential user device density. A mathematical model is also developed in order to easily obtain a measure of the potential capacity in high density scenarios.

Gene-environment interactions involving functional variants: Results from the Breast Cancer Association Consortium.

Investigating the most likely causal variants identified by fine‐mapping analyses may improve the power to detect gene–environment interactions. We assessed the interplay between 70 single nucleotide polymorphisms identified by genetic fine‐scale mapping of susceptibility loci and 11 epidemiological breast cancer risk factors in relation to breast cancer. Analyses were conducted on up to 58,573 subjects (26,968 cases and 31,605 controls) from the Breast Cancer Association Consortium, in one of the largest studies of its kind. Analyses were carried out separately for estrogen receptor (ER) positive (ER+) and ER negative (ER–) disease. The Bayesian False Discovery Probability (BFDP) was computed to assess the noteworthiness of the results. Four potential gene–environment interactions were identified as noteworthy (BFDP < 0.80) when assuming a true prior interaction probability of 0.01. The strongest interaction result in relation to overall breast cancer risk was found between CFLAR‐rs7558475 and current smoking (ORint = 0.77, 95% CI: 0.67–0.88, p int = 1.8 × 10−4). The interaction with the strongest statistical evidence was found between 5q14‐rs7707921 and alcohol consumption (ORint =1.36, 95% CI: 1.16–1.59, p int = 1.9 × 10−5) in relation to ER– disease risk. The remaining two gene–environment interactions were also identified in relation to ER– breast cancer risk and were found between 3p21‐rs6796502 and age at menarche (ORint = 1.26, 95% CI: 1.12–1.43, p int =1.8 × 10−4) and between 8q23‐rs13267382 and age at first full‐term pregnancy (ORint = 0.89, 95% CI: 0.83–0.95, p int = 5.2 × 10−4). While these results do not suggest any strong gene–environment interactions, our results may still be useful to inform experimental studies. These may in turn, shed light on the potential interactions observed.

A Bayesian multinomial logistic exposure model for estimating probabilities of competing sources of nest failure

Understanding causes of nest loss is critical for the management of endangered bird populations. Available methods for estimating nest loss probabilities to competing sources do not allow for random effects and covariation among sources, and there are few data simulation methods or goodness‐of‐fit (GOF) tests for such models. We developed a Bayesian multinomial extension of the widely used logistic exposure (LE) nest survival model which can incorporate multiple random effects and fixed‐effect covariates for each nest loss category. We investigated the performance of this model and the accompanyingGOFtest by analysing simulated nest fate datasets with and without age‐biased discovery probability, and by comparing the estimates with those of traditional fixed‐effects estimators. We then exemplify the use of the multinomialLEmodel andGOFtest by analysing Piping PloverCharadrius melodusnest fate data (n = 443) to explore the effects of wire cages (exclosures) constructed around nests, which are used to protect nests from predation but can lead to increased nest abandonment rates. Mean parameter estimates of the random‐effects multinomialLEmodel were all within 1 sd of the true values used to simulate the datasets. Age‐biased discovery probability did not result in biased parameter estimates. Traditional fixed‐effects models provided estimates with a high bias of up to 43% with a mean of 71% smaller standard deviations. TheGOFtest identified models that were a poor fit to the simulated data. For the Piping Plover dataset, the fixed‐effects model was less well‐supported than the random‐effects model and underestimated the risk of exclosure use by 16%. The random‐effects model estimated a range of 1–6% probability of abandonment for nests not protected by exclosures across sites and 5–41% probability of abandonment for nests with exclosures, suggesting that the magnitude of exclosure‐related abandonment is site‐specific. Our results demonstrate that unmodelled heterogeneity can result in biased estimates potentially leading to incorrect management recommendations. The Bayesian multinomialLEmodel offers a flexible method of incorporating random effects into an analysis of nest failure and is robust to age‐biased nest discovery probability. This model can be generalized to other staggered‐entry, time‐to‐hazard situations.