GECKO is a genetic algorithm to classify and explore high throughput sequencing data

Aubin Thomas,Julie Brooke,Jean-Philippe Villemin,Claudio Lorenzi,Gregory Beurier,Robert Sabatier,William Ritchie,Christelle Reynes,Lucile Broseus,Alban Mancheron,Sylvain Barriere

doi:10.1038/s42003-019-0456-9

Abstract

Comparative analysis of high throughput sequencing data between multiple conditions often involves mapping of sequencing reads to a reference and downstream bioinformatics analyses. Both of these steps may introduce heavy bias and potential data loss. This is especially true in studies where patient transcriptomes or genomes may vary from their references, such as in cancer. Here we describe a novel approach and associated software that makes use of advances in genetic algorithms and feature selection to comprehensively explore massive volumes of sequencing data to classify and discover new sequences of interest without a mapping step and without intensive use of specialized bioinformatics pipelines. We demonstrate that our approach called GECKO for GEnetic Classification using k-mer Optimization is effective at classifying and extracting meaningful sequences from multiple types of sequencing approaches including mRNA, microRNA, and DNA methylome data.

Highlights

Comparative analysis of high throughput sequencing data between multiple conditions often involves mapping of sequencing reads to a reference and downstream bioinformatics analyses
GECKO is designed around two main steps; these are a k-mer matrix preparation step and an adaptive genetic algorithm (Fig. 1)
GECKO will iterate through 20,000 generations or stop when the number of new solutions discovered throughout generations slows down

Summary

Introduction

Comparative analysis of high throughput sequencing data between multiple conditions often involves mapping of sequencing reads to a reference and downstream bioinformatics analyses. To remove the requirement of a reference, recent methodologies use k-mer representation; they directly compare the counts of nucleotide sequences of length k between samples[5] These approaches have been successful at detecting novel transcripts but only on a very small subset of RNA sequencing data[4] and would be impossible to implement for the classification of large patient cohorts using the entire transcriptome. Exploring a large set of k-mers to classify samples can be framed as a global optimization problem for which many recent approaches have been published and compared[8] Amongst these is a class of nature-inspired algorithms termed Genetic Algorithm which are based on the processes of mutation, crossing over and natural selection. By visualizing how the genetic algorithm evolves to find solutions, GECKO can be used to explore novel sequences or groups of functionally related sequences associated with normal biology and disease

Methods

Results

Conclusion