Dataset Information

Trainable high resolution melt curve machine learning classifier for large-scale reliable genotyping of sequence variants.

ABSTRACT: High resolution melt (HRM) is gaining considerable popularity as a simple and robust method for genotyping sequence variants. However, accurate genotyping of an unknown sample for which a large number of possible variants may exist will require an automated HRM curve identification method capable of comparing unknowns against a large cohort of known sequence variants. Herein, we describe a new method for automated HRM curve classification based on machine learning methods and learned tolerance for reaction condition deviations. We tested this method in silico through multiple cross-validations using curves generated from 9 different simulated experimental conditions to classify 92 known serotypes of Streptococcus pneumoniae and demonstrated over 99% accuracy with 8 training curves per serotype. In vitro verification of the algorithm was tested using sequence variants of a cancer-related gene and demonstrated 100% accuracy with 3 training curves per sequence variant. The machine learning algorithm enabled reliable, scalable, and automated HRM genotyping analysis with broad potential clinical and epidemiological applications.

SUBMITTER: Athamanolap P

PROVIDER: S-EPMC4183555 | biostudies-literature | 2014

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Trainable high resolution melt curve machine learning classifier for large-scale reliable genotyping of sequence variants.

Athamanolap Pornpat P Parekh Vishwa V Fraley Stephanie I SI Agarwal Vatsal V Shin Dong J DJ Jacobs Michael A MA Wang Tza-Huei TH Yang Samuel S

PloS one 20141002 9

High resolution melt (HRM) is gaining considerable popularity as a simple and robust method for genotyping sequence variants. However, accurate genotyping of an unknown sample for which a large number of possible variants may exist will require an automated HRM curve identification method capable of comparing unknowns against a large cohort of known sequence variants. Herein, we describe a new method for automated HRM curve classification based on machine learning methods and learned tolerance f ...[more]

PMID: 25275518

Similar Datasets

Project description:BACKGROUND: High-throughput sequencing technologies offer new perspectives for biomedical, agronomical and evolutionary research. Promising progresses now concern the application of these technologies to large-scale studies of genetic variation. Such studies require the genotyping of high numbers of samples. This is theoretically possible using 454 pyrosequencing, which generates billions of base pairs of sequence data. However several challenges arise: first in the attribution of each read produced to its original sample, and second, in bioinformatic analyses to distinguish true from artifactual sequence variation. This pilot study proposes a new application for the 454 GS FLX platform, allowing the individual genotyping of thousands of samples in one run. A probabilistic model has been developed to demonstrate the reliability of this method. RESULTS: DNA amplicons from 1,710 rodent samples were individually barcoded using a combination of tags located in forward and reverse primers. Amplicons consisted in 222 bp fragments corresponding to DRB exon 2, a highly polymorphic gene in mammals. A total of 221,789 reads were obtained, of which 153,349 were finally assigned to original samples. Rules based on a probabilistic model and a four-step procedure, were developed to validate sequences and provide a confidence level for each genotype. The method gave promising results, with the genotyping of DRB exon 2 sequences for 1,407 samples from 24 different rodent species and the sequencing of 392 variants in one half of a 454 run. Using replicates, we estimated that the reproducibility of genotyping reached 95%. CONCLUSIONS: This new approach is a promising alternative to classical methods involving electrophoresis-based techniques for variant separation and cloning-sequencing for sequence determination. The 454 system is less costly and time consuming and may enhance the reliability of genotypes obtained when high numbers of samples are studied. It opens up new perspectives for the study of evolutionary and functional genetics of highly polymorphic genes like major histocompatibility complex genes in vertebrates or loci regulating self-compatibility in plants. Important applications in biomedical research will include the detection of individual variation in disease susceptibility. Similarly, agronomy will benefit from this approach, through the study of genes implicated in productivity or disease susceptibility traits.

Dataset Information

Trainable high resolution melt curve machine learning classifier for large-scale reliable genotyping of sequence variants.

Publications

Trainable high resolution melt curve machine learning classifier for large-scale reliable genotyping of sequence variants.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets