Random forest classifier improving phenylketonuria screening performance in two Chinese populations.
Ontology highlight
ABSTRACT:
SUBMITTER: Song Y
PROVIDER: S-EPMC9592754 | biostudies-literature | 2022
REPOSITORIES: biostudies-literature
Publications
Random forest classifier improving phenylketonuria screening performance in two Chinese populations.
Song Yingnan Y Yin Zhe Z Zhang Chuan C Hao Shengju S Li Haibo H Wang Shifan S Yang Xiangchun X Li Qiong Q Zhuang Danyan D Zhang Xinyuan X Cao Zongfu Z Ma Xu X
Frontiers in molecular biosciences 20221011
Phenylketonuria (PKU) is a genetic disorder with amino acid metabolic defect, which does great harms to the development of newborns and children. Early diagnosis and treatment can effectively prevent the disease progression. Here we developed a PKU screening model using random forest classifier (RFC) to improve PKU screening performance with excellent sensitivity, false positive rate (FPR) and positive predictive value (PPV) in all the validation dataset and two testing Chinese populations. RFC ...[more]
Similar Datasets
Project description:BackgroundMethylmalonic acidemia (MMA) is one of the most common hereditary organic acid metabolism disorders that endangers the lives and health of infants and children. Early detection and intervention before the appearance of a newborn's clinical symptoms can control disease progression and prevent or mitigate its serious consequences.Methods42,004 newborns from two Chinese populations were included in the study. The small molecular metabolite analytes were detected from the dried blood spot (DBS) samples by MS/MS. Genetic analysis of 68 Chinese MMA cases were performed by whole-exome sequencing and Sanger sequencing. Random forest classifiers (RFC) were constructed to improve the MMA screening performance and genotype prediction in two Chinese populations. Meanwhile, other six machine learning models were trained to separate MMA patients from normal newborns. Model performance was assessed using accuracy, sensitivity, specificity, false positive rate (FPR), and positive predictive value (PPV) and the area under the receiver operating characteristic curve (AUC).ResultsIn the total 42,004 newborn samples, 68 MMA cases were identified by genetic analysis, 42 cases of which were caused by variants in MMACHC, 24 cases by variants in MMUT, and two cases by variants in MMAA. Three novel variants including c.449T>G (p.I150R) of MMACHC, c.1151C>T (p.S384F) and c.1091_1108delins (p.Y364Sfs*4) in MMUT were identified in the MMA patients. RFC for newborn screening of MMA performed best as compared to several other classification models based on machine learning with 100% sensitivity, low FPR, excellent PPV and AUC. In addition, the subdivision RFC for MMA genotype prediction was constructed with superior performance.ConclusionsIt can be seen that RFC is extremely helpful for detection and genotype prediction in the newborn MMA screening. In addition, our findings extend the variant spectrum of genes related to MMA.
| S-EPMC11552112 | biostudies-literature
Project description:We propose a random forest classifier for detecting rare variants from sequencing errors in Next Generation Sequencing (NGS) data from viral populations. The method utilizes counts of varying length of k-mers from the reads of a viral population to train a Random forest classifier, called MultiRes, that classifies k-mers as erroneous or rare variants. Our algorithm is rooted in concepts from signal processing and uses a frame-based representation of k-mers. Frames are sets of non-orthogonal basis functions that were traditionally used in signal processing for noise removal. We define discrete spatial signals for genomes and sequenced reads, and show that k-mers of a given size constitute a frame. We evaluate MultiRes on simulated and real viral population datasets, which consist of many low frequency variants, and compare it to the error detection methods used in correction tools known in the literature. MultiRes has 4 to 500 times less false positives k-mer predictions compared to other methods, essential for accurate estimation of viral population diversity and their de-novo assembly. It has high recall of the true k-mers, comparable to other error correction methods. MultiRes also has greater than 95% recall for detecting single nucleotide polymorphisms (SNPs) and fewer false positive SNPs, while detecting higher number of rare variants compared to other variant calling methods for viral populations. The software is available freely from the GitHub link https://github.com/raunaq-m/MultiRes.
| S-EPMC5548337 | biostudies-literature
Project description:Herein, we present the results of a machine learning approach we developed to single out correct 3D docking models of protein-protein complexes obtained by popular docking software. To this aim, we generated 3×104 docking models for each of the 230 complexes in the protein-protein benchmark, version 5, using three different docking programs (HADDOCK, FTDock and ZDOCK), for a cumulative set of ≈7×106 docking models. Three different machine learning approaches (Random Forest, Supported Vector Machine and Perceptron) were used to train classifiers with 158 different scoring functions (features). The Random Forest algorithm outperformed the other two algorithms and was selected for further optimization. Using a features selection algorithm, and optimizing the random forest hyperparameters, allowed us to train and validate a random forest classifier, named COnservation Driven Expert System (CoDES). Testing of CoDES on independent datasets, as well as results of its comparative performance with machine learning methods recently developed in the field for the scoring of docking decoys, confirm its state-of-the-art ability to discriminate correct from incorrect decoys both in terms of global parameters and in terms of decoys ranked at the top positions.Supplementary informationSupplementary data are available at Bioinformatics Advances online.Software and data availability statementThe docking models are available at https://doi.org/10.5281/zenodo.4012018. The programs underlying this article will be shared on request to the corresponding authors.
| S-EPMC9710594 | biostudies-literature
Project description:BackgroundWhile random forests are one of the most successful machine learning methods, it is necessary to optimize their performance for use with datasets resulting from a two-phase sampling design with a small number of cases-a common situation in biomedical studies, which often have rare outcomes and covariates whose measurement is resource-intensive.MethodsUsing an immunologic marker dataset from a phase III HIV vaccine efficacy trial, we seek to optimize random forest prediction performance using combinations of variable screening, class balancing, weighting, and hyperparameter tuning.ResultsOur experiments show that while class balancing helps improve random forest prediction performance when variable screening is not applied, class balancing has a negative impact on performance in the presence of variable screening. The impact of the weighting similarly depends on whether variable screening is applied. Hyperparameter tuning is ineffective in situations with small sample sizes. We further show that random forests under-perform generalized linear models for some subsets of markers, and prediction performance on this dataset can be improved by stacking random forests and generalized linear models trained on different subsets of predictors, and that the extent of improvement depends critically on the dissimilarities between candidate learner predictions.ConclusionIn small datasets from two-phase sampling design, variable screening and inverse sampling probability weighting are important for achieving good prediction performance of random forests. In addition, stacking random forests and simple linear models can offer improvements over random forests.
| S-EPMC8607560 | biostudies-literature
Project description:BackgroundIdentifying lncRNA-disease associations not only helps to better comprehend the underlying mechanisms of various human diseases at the lncRNA level but also speeds up the identification of potential biomarkers for disease diagnoses, treatments, prognoses, and drug response predictions. However, as the amount of archived biological data continues to grow, it has become increasingly difficult to detect potential human lncRNA-disease associations from these enormous biological datasets using traditional biological experimental methods. Consequently, developing new and effective computational methods to predict potential human lncRNA diseases is essential.ResultsUsing a combination of incremental principal component analysis (IPCA) and random forest (RF) algorithms and by integrating multiple similarity matrices, we propose a new algorithm (IPCARF) based on integrated machine learning technology for predicting lncRNA-disease associations. First, we used two different models to compute a semantic similarity matrix of diseases from a directed acyclic graph of diseases. Second, a characteristic vector for each lncRNA-disease pair is obtained by integrating disease similarity, lncRNA similarity, and Gaussian nuclear similarity. Then, the best feature subspace is obtained by applying IPCA to decrease the dimension of the original feature set. Finally, we train an RF model to predict potential lncRNA-disease associations. The experimental results show that the IPCARF algorithm effectively improves the AUC metric when predicting potential lncRNA-disease associations. Before the parameter optimization procedure, the AUC value predicted by the IPCARF algorithm under 10-fold cross-validation reached 0.8529; after selecting the optimal parameters using the grid search algorithm, the predicted AUC of the IPCARF algorithm reached 0.8611.ConclusionsWe compared IPCARF with the existing LRLSLDA, LRLSLDA-LNCSIM, TPGLDA, NPCMF, and ncPred prediction methods, which have shown excellent performance in predicting lncRNA-disease associations. The compared results of 10-fold cross-validation procedures show that the predictions of the IPCARF method are better than those of the other compared methods.
| S-EPMC8017839 | biostudies-literature
Project description:The development of new protein-ligand scoring functions using machine learning algorithms, such as random forest, has been of significant interest. By efficiently utilizing expanded feature sets and a large set of experimental data, random forest based scoring functions (RFbScore) can achieve better correlations to experimental protein-ligand binding data with known crystal structures; however, more extensive tests indicate that such enhancement in scoring power comes with significant under-performance in docking and screening power tests compared to traditional scoring functions. In this work, to improve scoring-docking-screening powers of protein-ligand docking functions simultaneously, we have introduced a Δvina RF parameterization and feature selection framework based on random forest. Our developed scoring function Δvina RF20 , which employs 20 descriptors in addition to the AutoDock Vina score, can achieve superior performance in all power tests of both CASF-2013 and CASF-2007 benchmarks compared to classical scoring functions. The Δvina RF20 scoring function and its code are freely available on the web at: https://www.nyu.edu/projects/yzhang/DeltaVina. © 2016 Wiley Periodicals, Inc.
| S-EPMC5140681 | biostudies-literature
Project description:Efficiency, memory consumption, and robustness are common problems with many popular methods for data analysis. As a solution, we present Random Bits Forest (RBF), a classification and regression algorithm that integrates neural networks (for depth), boosting (for width), and random forests (for prediction accuracy). Through a gradient boosting scheme, it first generates and selects ~10,000 small, 3-layer random neural networks. These networks are then fed into a modified random forest algorithm to obtain predictions. Testing with datasets from the UCI (University of California, Irvine) Machine Learning Repository shows that RBF outperforms other popular methods in both accuracy and robustness, especially with large datasets (N > 1000). The algorithm also performed highly in testing with an independent data set, a real psoriasis genome-wide association study (GWAS).
| S-EPMC4957112 | biostudies-literature
Project description:MicroRNAs are key regulators of eukaryotic gene expression whose fundamental role has already been identified in many cell pathways. The correct identification of miRNAs targets is still a major challenge in bioinformatics and has motivated the development of several computational methods to overcome inherent limitations of experimental analysis. Indeed, the best results reported so far in terms of specificity and sensitivity are associated to machine learning-based methods for microRNA-target prediction. Following this trend, in the current paper we discuss and explore a microRNA-target prediction method based on a random forest classifier, namely RFMirTarget. Despite its well-known robustness regarding general classifying tasks, to the best of our knowledge, random forest have not been deeply explored for the specific context of predicting microRNAs targets. Our framework first analyzes alignments between candidate microRNA-target pairs and extracts a set of structural, thermodynamics, alignment, seed and position-based features, upon which classification is performed. Experiments have shown that RFMirTarget outperforms several well-known classifiers with statistical significance, and that its performance is not impaired by the class imbalance problem or features correlation. Moreover, comparing it against other algorithms for microRNA target prediction using independent test data sets from TarBase and starBase, we observe a very promising performance, with higher sensitivity in relation to other methods. Finally, tests performed with RFMirTarget show the benefits of feature selection even for a classifier with embedded feature importance analysis, and the consistency between relevant features identified and important biological properties for effective microRNA-target gene alignment.
| S-EPMC3724815 | biostudies-literature
Project description:BackgroundMicrosatellite instability (MSI) is a common genomic alteration in colorectal cancer, endometrial carcinoma, and other solid tumors. MSI is characterized by a high degree of polymorphism in microsatellite lengths owing to the deficiency in the mismatch repair system. Based on the degree, MSI can be classified as microsatellite instability-high (MSI-H) and microsatellite stable (MSS). MSI is a predictive biomarker for immunotherapy efficacy in advanced/metastatic solid tumors, especially in colorectal cancer patients. Several computational approaches based on target panel sequencing data have been used to detect MSI; however, they are considerably affected by the sequencing depth and panel size.ResultsWe developed MSIFinder, a python package for automatic MSI classification, using random forest classifier (RFC)-based genome sequencing, which is a machine learning technology. We included 19 MSI-H and 25 MSS samples as training sets. First, we selected 54 feature markers from the training sets, built an RFC model, and validated the classifier using a test set comprising 21 MSI-H and 379 MSS samples. With this test set, MSIFinder achieved a sensitivity (recall) of 1.0, a specificity of 0.997, an accuracy of 0.998, a positive predictive value of 0.954, an F1 score of 0.977, and an area under the curve of 0.999. To further verify the robustness and effectiveness of the model, we used a prospective cohort consisting of 18 MSI-H samples and 122 MSS samples. MSIFinder achieved a sensitivity (recall) of 1.0 and a specificity of 1.0. We discovered that MSIFinder is less affected by a low sequencing depth and can achieve a concordance of 0.993 while exhibiting a sequencing depth of 100×. Furthermore, we realized that MSIFinder is less affected by the panel size and can achieve a concordance of 0.99 when the panel size is 0.5 M (million bases).ConclusionThese results indicate that MSIFinder is a robust and effective MSI classification tool that can provide reliable MSI detection for scientific and clinical purposes.
| S-EPMC8042960 | biostudies-literature
Project description:Ovarian cancer (OC) is associated with a poor prognosis due to difficulties in early detection. The aims of the present study were to construct a recurrence risk prediction model and to reveal important OC genes or pathways. RNA sequencing data was obtained for 307 OC samples, and the corresponding clinical data were downloaded from The Cancer Genome Atlas database. Additionally, two validation datasets, GSE44104 (20 recurrent and 40 non?recurrent OC samples) and GSE49997 (204 OC samples), were obtained from the Gene Expression Omnibus database. Differentially expressed genes were screened using the differential expression via distance synthesis algorithm, followed by gene ontology enrichment analysis and weighted gene coexpression network analysis (WGCNA). Furthermore, subnetwork analysis was conducted for the protein?protein interaction (PPI) network using the BioNet package. Finally, a random forest classifier was constructed based on the subnetwork nodes, and its reliability was validated using the GSE44104 and GSE49997 validation datasets. A total of 44 upregulated and 117 downregulated genes were identified in the recurrent samples. Enrichment analysis indicated that cytochrome P450 family 17 subfamily A member 1 (CYP17A1) was associated with 'positive regulation of steroid hormone biosynthetic processes'. WGCNA identified turquoise and grey modules that were significantly correlated with status and prognosis. A significant PPI subnetwork containing 16 nodes was also identified, including: Transcription factor GATA?4; fibroblast growth factor 9; aromatase; 3??hydroxysteroid dehydrogenase/?5?4?isomerase type 2; corticosteroid 11??dehydrogenase isozyme 1; CYP17A1; pituitary homeobox 2; left?right determination factor 1; homeobox protein ARX; estrogen receptor ?; steroidogenic factor 1; forkhead box protein L2; myocardin; steroidogenic acute regulatory protein mitochondrial; vesicular inhibitory amino acid transporter; and twist?related protein 1. A random forest classifier was constructed using the subnetwork nodes as feature genes, which exhibited a 92% true positive rate when classifying recurrent and non?recurrent OC samples. The classifying efficiency of the random forest classifier was validated using the two other independent datasets. Overall, 44 upregulated and 117 downregulated genes associated with OC recurrence were identified. Furthermore, the 16 subnetwork node genes that were identified may be important molecules in OC recurrence.
| S-EPMC6102638 | biostudies-literature