Benchmarking of computational methods for predicting circRNA-disease associations.

Accumulating evidences demonstrate that circular RNA (circRNA) plays an important role in human diseases. Identification of circRNA-disease associations can help for the diagnosis of human diseases, while the traditional method based on biological experiments is time-consuming. In order to address the limitation, a series of computational methods have been proposed in recent years. However, few works have summarized these methods or compared the performance of them. In this paper, we divided the existing methods into three categories: information propagation, traditional machine learning and deep learning. Then, the baseline methods in each category are introduced in detail. Further, 5 different datasets are collected, and 14 representative methods of each category are selected and compared in the 5-fold, 10-fold cross-validation and the de novo experiment. In order to further evaluate the effectiveness of these methods, six common cancers are selected to compare the number of correctly identified circRNA-disease associations in the top-10, top-20, top-50, top-100 and top-200. In addition, according to the results, the observation about the robustness and the character of these methods are concluded. Finally, the future directions and challenges are discussed.

A granularity-level information fusion strategy on hypergraph transformer for predicting synergistic effects of anticancer drugs.

Combination therapy has exhibited substantial potential compared to monotherapy. However, due to the explosive growth in the number of cancer drugs, the screening of synergistic drug combinations has become both expensive and time-consuming. Synergistic drug combinations refer to the concurrent use of two or more drugs to enhance treatment efficacy. Currently, numerous computational methods have been developed to predict the synergistic effects of anticancer drugs. However, there has been insufficient exploration of how to mine drug and cell line data at different granularity levels for predicting synergistic anticancer drug combinations. Therefore, this study proposes a granularity-level information fusion strategy based on the hypergraph transformer, named HypertranSynergy, to predict synergistic effects of anticancer drugs. HypertranSynergy introduces synergistic connections between cancer cell lines and drug combinations using hypergraph. Then, the Coarse-grained Information Extraction (CIE) module merges the hypergraph with a transformer for node embeddings. In the CIE module, Contranorm is a normalization layer that mitigates over-smoothing, while Gaussian noise addresses local information gaps. Additionally, the Fine-grained Information Extraction (FIE) module assesses fine-grained information's impact on predictions by employing similarity-aware matrices from drug/cell line features. Both CIE and FIE modules are integrated into HypertranSynergy. In addition, HypertranSynergy achieved the AUC of 0.93${\pm }$0.01 and the AUPR of 0.69${\pm }$0.02 in 5-fold cross-validation of classification task, and the RMSE of 13.77${\pm }$0.07 and the PCC of 0.81${\pm }$0.02 in 5-fold cross-validation of regression task. These results are better than most of the state-of-the-art models.

HiC4D: forecasting spatiotemporal Hi-C data with residual ConvLSTM.

The Hi-C experiments have been extensively used for the studies of genomic structures. In the last few years, spatiotemporal Hi-C has largely contributed to the investigation of genome dynamic reorganization. However, computationally modeling and forecasting spatiotemporal Hi-C data still have not been seen in the literature. We present HiC4D for dealing with the problem of forecasting spatiotemporal Hi-C data. We designed and benchmarked a novel network and named it residual ConvLSTM (ResConvLSTM), which is a combination of residual network and convolutional long short-term memory (ConvLSTM). We evaluated our new ResConvLSTM networks and compared them with the other five methods, including a naïve network (NaiveNet) that we designed as a baseline method and four outstanding video-prediction methods from the literature: ConvLSTM, spatiotemporal LSTM (ST-LSTM), self-attention LSTM (SA-LSTM) and simple video prediction (SimVP). We used eight different spatiotemporal Hi-C datasets for the blind test, including two from mouse embryogenesis, one from somatic cell nuclear transfer (SCNT) embryos, three embryogenesis datasets from different species and two non-embryogenesis datasets. Our evaluation results indicate that our ResConvLSTM networks almost always outperform the other methods on the eight blind-test datasets in terms of accurately predicting the Hi-C contact matrices at future time-steps. Our benchmarks also indicate that all of the methods that we benchmarked can successfully recover the boundaries of topologically associating domains called on the experimental Hi-C contact matrices. Taken together, our benchmarks suggest that HiC4D is an effective tool for predicting spatiotemporal Hi-C data. HiC4D is publicly available at both http://dna.cs.miami.edu/HiC4D/ and https://github.com/zwang-bioinformatics/HiC4D/.

Predicting network dynamics without requiring the knowledge of the interaction graph.

A network consists of two interdependent parts: the network topology or graph, consisting of the links between nodes and the network dynamics, specified by some governing equations. A crucial challenge is the prediction of dynamics on networks, such as forecasting the spread of an infectious disease on a human contact network. Unfortunately, an accurate prediction of the dynamics seems hardly feasible, because the network is often complicated and unknown. In this work, given past observations of the dynamics on a fixed graph, we show the contrary: Even without knowing the network topology, we can predict the dynamics. Specifically, for a general class of deterministic governing equations, we propose a two-step prediction algorithm. First, we obtain a surrogate network by fitting past observations of every nodal state to the dynamical model. Second, we iterate the governing equations on the surrogate network to predict the dynamics. Surprisingly, even though there is no similarity between the surrogate topology and the true topology, the predictions are accurate, for a considerable prediction time horizon, for a broad range of observation times, and in the presence of a reasonable noise level. The true topology is not needed for predicting dynamics on networks, since the dynamics evolve in a subspace of astonishingly low dimension compared to the size and heterogeneity of the graph. Our results constitute a fresh perspective on the broad field of nonlinear dynamics on complex networks.

Brain fingerprinting and cognitive behavior predicting using functional connectome of high inter-subject variability.

The functional connectivity (FC) graph of the brain has been widely recognized as a ``fingerprint'' that can be used to identify individuals from a group of subjects. Research has indicated that individual identification accuracy can be improved by eliminating the impact of shared information among individuals. However, current research extracts not only shared information of inter-subject but also individual-specific information from FC graphs, resulting in incomplete separation of shared information and fingerprint information among individuals, leading to lower individual identification accuracy across all functional magnetic resonance imaging (fMRI) states session pairs and poor cognitive behavior prediction performance. In this paper, we propose a method to enhance inter-subject variability combining conditional variational autoencoder (CVAE) network and sparse dictionary learning (SDL) module. By embedding fMRI state information in the encoding and decoding processes, the CVAE network can better capture and represent the common features among individuals and enhance inter-subject variability by residual. Our experimental results on Human Connectome Project (HCP) data show that the refined connectomes obtained by using CVAE with SDL can accurately distinguish an individual from the remaining participants. The success accuracies reached 99.7 % and 99.6 % in the session pair rest1-rest2 and reverse rest2-rest1, respectively. In the identification experiment involving task-task combinations carried out on the same day, the identification accuracies ranged from 94.2 % to 98.8 %. Furthermore, we showed the Frontoparietal and Default networks make the most significant contributions to individual identification and the edges that significantly contribute to individual identification are found within and between the Frontoparietal and Default networks. Additionally, high-level cognitive behaviors can also be better predicted with the obtained refined connectomes, suggesting that higher fingerprinting can be useful for resulting in higher behavioral associations. In summary, our proposed framework provides a promising approach to use functional connectivity networks for studying cognition and behavior, promoting a deeper understanding of brain functions.

Specific serum autoantibodies predict the development and progression of Alzheimer's disease with high accuracy.

Autoimmunity plays a key role in the pathogenesis of Alzheimer's disease (AD). However, whether autoantibodies in peripheral blood can be used as biomarkers for AD has been elusive. Serum samples were obtained from 1,686 participants, including 767 with AD, 146 with mild cognitive impairment (MCI), 255 with other neurodegenerative diseases, and 518 healthy controls. Specific autoantibodies were measured using a custom-made immunoassay. Multivariate support vector machine models were employed to investigate the correlation between serum autoantibody levels and disease states. As a result, seven candidate AD-specific autoantibodies were identified, including MAPT, DNAJC8, KDM4D, SERF1A, CDKN1A, AGER, and ASXL1. A classification model with high accuracy (area under the curve (AUC) = 0.94) was established. Importantly, these autoantibodies could distinguish AD from other neurodegenerative diseases and out-performed amyloid and tau protein concentrations in cerebrospinal fluid in predicting cognitive decline (P < 0.001). This study indicated that AD onset and progression are possibly accompanied by an unappreciated serum autoantibody response. Therefore, future studies could optimize its application as a convenient biomarker for the early detection of AD.

A new score for predicting intracranial hemorrhage in patients using antiplatelet drugs.

Antiplatelet drugs in patients increase the risk of intracranial hemorrhage (ICH), which can seriously affect patients' quality of life and even endanger their lives. Currently, there is no specific score for predicting the risk of ICH caused by antiplatelet drugs. We aimed to identify factors associated with ICH in patients on antiplatelet drugs and to construct and validate a predictive model that would provide a validated tool for the clinic. Data were obtained from the patient medical records inpatient system. Prediction models were built by logistic regression, the area under the curve (AUC), and column line plots. Internal validation, analytical identification and calibration of the model using AUC, calibration curves and Hosmer-Lemeshow test. The registration number of this study is ChiCTR2000031909, and the ethical review number is 2020KY087. This single-center retrospective study enrolled 753 patients treated with antiplatelet drugs, including 527 in the development cohort. Multifactorial analysis showed that male, headache or vomiting, hypertension, cerebrovascular disease, CT-defined white matter hypodensity, abnormal GCS, fibrinogen and D-dimer were independent risk factors for ICH, and lipid-lowering drugs was a protective factor. The model was constructed using these nine factors with an AUC value of 0.949. In the validation cohort, the model showed good discriminatory power with an AUC value of 0.943 and good calibration (Hosmer-Lemeshow test P value of 0.818). Based on 9 factors, we derived and validated a predictive model for ICH with antiplatelet drugs in patients. The model has good predictive value and may be an effective tool to reduce the occurrence of ICH.

Comprehensive analysis of single cell and bulk RNA sequencing reveals the heterogeneity of melanoma tumor microenvironment and predicts the response of immunotherapy.

BACKGROUND: Tumor microenvironment (TME) heterogeneity is an important factor affecting the treatment response of immune checkpoint inhibitors (ICI). However, the TME heterogeneity of melanoma is still widely characterized. METHODS: We downloaded the single-cell sequencing data sets of two melanoma patients from the GEO database, and used the "Scissor" algorithm and the "BayesPrism" algorithm to comprehensively analyze the characteristics of microenvironment cells based on single-cell and bulk RNA-seq data. The prediction model of immunotherapy response was constructed by machine learning and verified in three cohorts of GEO database. RESULTS: We identified seven cell types. In the Scissor+ subtype cell population, the top three were T cells, B cells and melanoma cells. In the Scissor- subtype, there are more macrophages. By quantifying the characteristics of TME, significant differences in B cells between responders and non-responders were observed. The higher the proportion of B cells, the better the prognosis. At the same time, macrophages in the non-responsive group increased significantly. Finally, nine gene features for predicting ICI response were constructed, and their predictive performance was superior in three external validation groups. CONCLUSION: Our study revealed the heterogeneity of melanoma TME and found a new predictive biomarker, which provided theoretical support and new insights for precise immunotherapy of melanoma patients.

A time series driven model for early sepsis prediction based on transformer module.

Sepsis remains a critical concern in intensive care units due to its high mortality rate. Early identification and intervention are paramount to improving patient outcomes. In this study, we have proposed predictive models for early sepsis prediction based on time-series data, utilizing both CNN-Transformer and LSTM-Transformer architectures. By collecting time-series data from patients at 4, 8, and 12 h prior to sepsis diagnosis and subjecting it to various network models for analysis and comparison. In contrast to traditional recurrent neural networks, our model exhibited a substantial improvement of approximately 20%. On average, our model demonstrated an accuracy of 0.964 (± 0.018), a precision of 0.956 (± 0.012), a recall of 0.967 (± 0.012), and an F1 score of 0.959 (± 0.014). Furthermore, by adjusting the time window, it was observed that the Transformer-based model demonstrated exceptional predictive capabilities, particularly within the earlier time window (i.e., 12 h before onset), thus holding significant promise for early clinical diagnosis and intervention. Besides, we employed the SHAP algorithm to visualize the weight distribution of different features, enhancing the interpretability of our model and facilitating early clinical diagnosis and intervention.

The role of neutrophil-to-lymphocyte ratio, platelet-to-lymphocyte ratio, and systemic immune inflammation index in predicting the necessity for surgery and therapeutic surgery in patients with anterior abdominal stab wounds.

BACKGROUND: In this diagnostic accuracy study, we examined the effectiveness of neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), and systemic immune inflammation index (SII) in predicting the need for surgical intervention in patients with anterior abdominal stab wounds (AASW) who exhibit unclear findings on physical examination yet remain hemodynamically stable. METHODS: Over a 7-year period, patients with AASW were retrospectively analyzed. Patients were divided into two groups as surgical (SG) and nonsurgical group (nSG). The SG were also divided into two groups as therapeutic surgery (TS) group and the non-therapeutic surgery (nTS) group. The groups were compared in terms of NLR, PLR values and SII scores. RESULTS: In a retrospective analysis of 199 patients with AASW, NLR, PLR and SII obtained during clinical follow-up of patients with AASW in whom the necessity for immediate surgery was unclear significantly predicted therapeutic surgery (p < 0.001 for all). These parameters did not show a significant difference in predicting the need for surgery at the admission. NLR showed an AUC of 0.971 and performed significantly better than PLR and SII (AUC = 0.874 and 0.902, respectively) in predicting TS. The optimal cut-off value for NLR was 3.33, with a sensitivity of 98.2%, a specificity of 90%, and a negative likelihood ratio of 0.02. Time from admission to surgery was significantly shorter in the TS group (p = 0.001). CONCLUSION: NLR, PLR and SII values may be useful in predicting therapeutic surgery during clinical follow-up in AASW patients with unclear physical examination findings and in whom immediate surgical decisions cannot be made.

Association between pathologic complete response and biochemical indicators after neoadjuvant therapy for HER2-positive breast cancer.

PURPOSE: This study investigated the changes in the fasting blood glucose (FBG), fasting triglyceride (FTG), and fasting total cholesterol (FTC) levels during neoadjuvant therapy (NAT) for human epidermal growth factor receptor 2 (HER2)-positive breast cancer (BC) and the association with pathologic complete response (pCR). METHODS: Relevant data from Sichuan Cancer Hospital from June 2019 to June 2022 were collected and analyzed, and FBG, FTG, and FTC were divided into baseline, change, and process groups, which were grouped to analyze the changes after receiving NAT and the association with pCR. RESULTS: In the estrogen receptor (ER)-negative subgroup, patients with low levels of FTG in the process group were more likely to achieve pCR compared to high levels, and in the progesterone receptor (PR)-negative subgroup, patients with lower FTG compared to higher FTG after receiving NAT was more likely to achieve pCR. CONCLUSIONS: Patients with HER2-positive BC undergoing NAT develop varying degrees of abnormalities (elevated or decreased) in FBG, FTG, and FTC; moreover, the status of FTG levels during NAT may predict pCR in ER-negative or PR-negative HER2-positive BC.Early monitoring and timely intervention for FTG abnormalities may enable this subset of patients to increase the likelihood of obtaining a pCR along with management of abnormal markers.

Does cleavage stage morphology increase the discriminatory power of prediction in blastocyst transfer outcome?

PURPOSE: To evaluate the contribution of the cleavage stage morphological parameters to the prediction of blastocyst transfer outcomes. METHODS: A retrospective study was conducted on 8383 single-blastocyst transfer cycles including 2246 fresh and 6137 vitrified-warmed cycles. XGboost, LASSO, and GLM algorithms were employed to establish models for assessing the predictive value of the cleavage stage morphological parameters in transfer outcomes. Four models were developed using each algorithm: all-in model with or without day 3 morphology and embryo quality-only model with or without day 3 morphology. RESULTS: The live birth rate was 48.04% in the overall cohort. The AUCs of the models with the algorithm of XGboost were 0.83, 0.82, 0.63, and 0.60; with LASSO were 0.66, 0.66, 0.61, and 0.60; and with GLM were 0.66, 0.66, 0.61, and 0.60 respectively. In models 1 and 2, female age, basal FSH, peak E2, endometrial thickness, and female BMI were the top five critical features for predicting live birth; In models 3 and 4, the most crucial factor was blastocyst formation on D5 rather than D6. In model 3, incorporating cleavage stage morphology, including early cleavage, D3 cell number, and fragmentation, was significantly associated with successful live birth. Additionally, the live birth rates for blastocysts derived from on-time, slow, and fast D3 embryos were 49.7%, 39.5%, and 52%, respectively. CONCLUSIONS: The value of cleavage stage morphological parameters in predicting the live birth outcome of single blastocyst transfer is limited.

Study on the Impact of Building Energy Predictions Considering Weather Errors of Neighboring Weather Stations.

Weather data errors affect energy management by influencing the accuracy of building energy predictions. This study presents a long short-term memory (LSTM) prediction model based on the "Energy Detective" dataset (Shanghai, China) and neighboring weather station data. The study analyzes the errors of different weather data sources (Detective and A) at the same latitude and longitude. Subsequently, it discusses the effects of weather errors from neighboring weather stations (Detective, A, B, C, and D) on energy forecasts for the next hour and day including the selection process for neighboring weather stations. Furthermore, it compares the forecast results for summer and autumn. The findings indicate a correlation between weather errors from neighboring weather stations and energy consumption. The median R-Square for predicting the next hour reached 0.95. The model's predictions for the next day exhibit a higher Prediction Interval Mean Width (139.0 in summer and 146.1 in autumn), indicating a greater uncertainty.

Development and validation of a nomogram for predicting high-burnout risk in nurses.

AIM: To develop a predictive model for high-burnout of nurses. DESIGN: A cross-sectional study. METHODS: This study was conducted using an online survey. Data were collected by the Chinese Maslach Burnout Inventory-General Survey (CMBI-GS) and self-administered questionnaires that included demographic, behavioural, health-related, and occupational variables. Participants were randomly divided into a development set and a validation set. In the development set, multivariate logistic regression analysis was conducted to identify factors associated with high-burnout risk, and a nomogram was constructed based on significant contributing factors. The discrimination, calibration, and clinical practicability of the nomogram were evaluated in both the development and validation sets using receiver operating characteristic (ROC) curve analysis, Hosmer-Lemeshow test, and decision curve analysis, respectively. Data analysis was performed using Stata 16.0 software. RESULTS: A total of 2750 nurses from 23 provinces of mainland China responded, with 1925 participants (70%) in a development set and 825 participants (30%) in a validation set. Workplace violence, shift work, working time per week, depression, stress, self-reported health, and drinking were significant contributors to high-burnout risk and a nomogram was developed using these factors. The ROC curve analysis demonstrated that the area under the curve of the model was 0.808 in the development set and 0.790 in the validation set. The nomogram demonstrated a high net benefit in the clinical decision curve in both sets. CONCLUSION: This study has developed and validated a predictive nomogram for identifying high-burnout in nurses. RELEVANCE TO CLINICAL PRACTICE: The nomogram conducted by our study will assist nursing managers in identifying at-high-risk nurses and understanding related factors, helping them implement interventions early and purposefully. REPORTING METHOD: The study adhered to the relevant EQUATOR reporting guidelines: TRIPOD Checklist for Prediction Model Development and Validation. PATIENT OR PUBLIC CONTRIBUTION: No patient or public contribution.

Random forest models of food safety behavior during the COVID-19 pandemic.

Machine learning approaches are increasingly being adopted as data analysis tools in scientific behavioral predictions. This paper utilizes a machine learning approach, Random Forest Model, to determine the top prediction variables of food safety behavioral changes during the pandemic. Data was collected among U.S. consumers on risk perception of COVID-19 and foodborne illness (FBI), food safety practice behaviors and demographics through online surveys at ten different time points from April 2020 through to May 2021; and post pandemic in May 2022. Random forest model was used to predict 14 food safety-related behaviors. The models for predicting Handwashing before cooking and Handwashing after eating had a good performance, with F-1 score of 0.93 and 0.88, respectively. Attitudes- related variables were determined to be important in predicting food safety behaviors. The importance ranking of the predicting variables were found to be changing over time.

Prediction of liver metastasis and recommended optimal follow-up nursing in rectal cancer.

We aimed to analyze and investigate the clinical factors that influence the occurrence of liver metastasis in locally advanced rectal cancer patients, with an attempt to assist patients in devising the optimal imaging-based follow-up nursing. Between June 2011 and May 2021, patients with rectal cancer at our hospital were retrospectively analyzed. A random survival forest model was developed to predict the probability of liver metastasis and provide a practical risk-based approach to surveillance. The results indicated that age, perineural invasion, and tumor deposit were significant factors associated with the liver metastasis and survival. The liver metastasis risk of the low-risk group was higher at 6-21 months, with a peak occurrence time in the 15th month. The liver metastasis risk of the high-risk group was higher at 0-24 months, with a peak occurrence time in the 8th month. In general, our clinical model could predict liver metastasis in rectal cancer patients. It provides a visualization tool that can aid physicians and nurses in making clinical decisions, by detecting the probability of liver metastasis.

Possibilities of predicting adverse cardiovascular events based on the analysis of clinical and instrumental research methods, as well as sST2 in patients after myocardial infarction.

OBJECTIVE: Aim: To determine the possibility of predicting adverse cardiovascular events based on the analysis of clinical and instrumental research methods, as well as sST2 in patients after myocardial infarction. PATIENTS AND METHODS: Materials and Methods: The study included 64 patients who suffered an acute myocardial infarction and underwent PCI with balloon angioplasty and stenting of the infarct-related vessel in the acute period. The predictors of adverse cardiovascular events were assessed events during 1 year of observation. Indicators of echocardiography and coronary angiography were assessed and concentrations sST2. RESULTS: Results: A worse prognosis was associated with intermediate ejection fraction (EF) (odds ratio (OR)=3.981, p<0.05), left aneurysm ventricle (LV) (OR=29.5, p<0.05), high concentrations of sST2 (OR=1.017, p<0.05) and scores on the Syntax scale (OR=1.001, p<0.05). CONCLUSION: Conclusions: In patients who underwent percutaneous coronary intervention for myocardial infarction, adverse outcome during the next 2 years is associated with coronary and echocardiographic parameters, as well as biochemical indicators of myocardial stress and fibrosis. HF patients with intermediate EF, LV aneurysm, high sST2 concentrations, and high Syntax scores have the worst prognosis.

A near-deterministic mutational hotspot in Pseudomonas fluorescens is constructed by multiple interacting genomic features.

Mutation - whilst stochastic - is frequently biased toward certain loci. When combined with selection this results in highly repeatable and predictable evolutionary outcomes. Immotile variants of the bacterium Pseudomonas fluorescens (SBW25) possess a 'mutational hotspot' that facilitates repeated occurrences of an identical de novo single nucleotide polymorphism when re-evolving motility, where ≥95% independent lines fix the mutation ntrB A289C. Identifying hotspots of similar potency in other genes and genomic backgrounds would prove valuable for predictive evolutionary models, but to do so we must understand the genomic features that enable such a hotspot to form. Here we reveal that genomic location, local nucleotide sequence, gene strandedness and presence of mismatch repair proteins operate in combination to facilitate the formation of this mutational hotspot. Our study therefore provides a framework for utilising genomic features to predict and identify hotspot positions capable of enforcing near-deterministic evolution.

Mutational hotspots lead to robust but suboptimal adaptive outcomes in certain environments.

The observed mutational spectrum of adaptive outcomes can be constrained by many factors. For example, mutational biases can narrow the observed spectrum by increasing the rate of mutation at isolated sites in the genome. In contrast, complex environments can shift the observed spectrum by defining fitness consequences of mutational routes. We investigate the impact of different nutrient environments on the evolution of motility in Pseudomonas fluorescens Pf0-2x (an engineered non-motile derivative of Pf0-1) in the presence and absence of a strong mutational hotspot. Previous work has shown that this mutational hotspot can be built and broken via six silent mutations, which provide rapid access to a mutation that rescues swimming motility and confers the strongest swimming phenotype in specific environments. Here, we evolved a hotspot and non-hotspot variant strain of Pf0-2x for motility under nutrient-rich (LB) and nutrient-limiting (M9) environmental conditions. We observed the hotspot strain consistently evolved faster across all environmental conditions and its mutational spectrum was robust to environmental differences. However, the non-hotspot strain had a distinct mutational spectrum that changed depending on the nutrient environment. Interestingly, while alternative adaptive mutations in nutrient-rich environments were equal to, or less effective than, the hotspot mutation, the majority of these mutations in nutrient-limited conditions produced superior swimmers. Our competition experiments mirrored these findings, underscoring the role of environment in defining both the mutational spectrum and the associated phenotype strength. This indicates that while mutational hotspots working in concert with natural selection can speed up access to robust adaptive mutations (which can provide a competitive advantage in evolving populations), they can limit exploration of the mutational landscape, restricting access to potentially stronger phenotypes in specific environments.

Distribution of mutation rates challenges evolutionary predictability.

Natural selection is commonly assumed to act on extensive standing genetic variation. Yet, accumulating evidence highlights the role of mutational processes creating this genetic variation: to become evolutionarily successful, adaptive mutants must not only reach fixation, but also emerge in the first place, i.e. have a high enough mutation rate. Here, we use numerical simulations to investigate how mutational biases impact our ability to observe rare mutational pathways in the laboratory and to predict outcomes in experimental evolution. We show that unevenness in the rates at which mutational pathways produce adaptive mutants means that most experimental studies lack power to directly observe the full range of adaptive mutations. Modelling mutation rates as a distribution, we show that a substantially larger target size ensures that a pathway mutates more commonly. Therefore, we predict that commonly mutated pathways are conserved between closely related species, but not rarely mutated pathways. This approach formalizes our proposal that most mutations have a lower mutation rate than the average mutation rate measured experimentally. We suggest that the extent of genetic variation is overestimated when based on the average mutation rate.