Distinguishing Species Using GC Contents in Mixed DNA or RNA Sequences

With the advent of whole transcriptome and genome analysis methods, classifying samples containing multiple origins has become a significant task. Nucleotide sequences can be allocated to a genome or transcriptome by aligning sequences to multiple target sequence sets, but this approach requires ext...

Full description

Saved in:
Bibliographic Details
Published in:Evolutionary Bioinformatics 2018, Vol.14, p.1176934318788866-1176934318788866
Main Authors: Karimi, Kamran, Wuitchik, Daniel M, Oldach, Matthew J, Vize, Peter D
Format: Article
Language:English
Subjects:
Accuracy
Authorship
Classification
Computer applications
Contaminants
Data analysis
Decision trees
Deoxyribonucleic acid
DNA
Endosymbionts
Experiments
Gene expression
Gene sequencing
Genera
Genomes
Methods
Next-generation sequencing
Nucleotide sequence
Nucleotides
Outliers (statistics)
Ribonucleic acid
RNA
Software or Database Review
Species
Support vector machines
Symbiodinium
Symbionts
Zooxanthellae
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:With the advent of whole transcriptome and genome analysis methods, classifying samples containing multiple origins has become a significant task. Nucleotide sequences can be allocated to a genome or transcriptome by aligning sequences to multiple target sequence sets, but this approach requires extensive computational resources and also depends on target sequence sets lacking contaminants, which is often not the case. Here, we demonstrate that raw sequences can be rapidly sorted into groups, in practice corresponding to genera, by exploiting differences in nucleotide GC content. To do so, we introduce GCSpeciesSorter, which uses classification, specifically Support Vector Machines (SVM) and the C4.5 decision tree generator, to differentiate sequences. It also implements a secondary BLAST feature to identify known outliers. In the test case presented, a hermatypic coral holobiont, the cnidarian host includes various endosymbionts. The best characterized and most common of these symbionts are zooxanthellae of the genus Symbiodinium. GCSpeciesSorter separates cnidarian from Symbiodinium sequences with a high degree of accuracy. We show that if the GC contents of the species differ enough, this method can be used to accurately distinguish the sequences of different species when using high-throughput sequencing technologies.
ISSN:1176-9343
1176-9343
DOI:10.1177/1176934318788866