Read Simulator

Usage

java -jar readSimulator.jar 
     -length <readLength>
     -frlength <fragmentLength>
     -SD <standardDeviation>
     -mutationrate <mutationRate (1.0 == 1%)>
     -gtf <"./inputFiles/Homo_sapiens.GRCh37.75.gtf">
     -fasta <"./inputFiles/Homo_sapiens.GRCh37.75.dna.toplevel.fa">
     -fidx <"./inputFiles/Homo_sapiens.GRCh37.75.dna.toplevel.fa.fai">
     -readcounts <"./inputFiles/readcounts.simulation">
     -od <"outputDir">
     [-debug]
     [-transciptome <"Homo_sapiens.GRCh37.75.cdna.all.fa">]

Background

In bioinformatics, Sequencing is a term used to describe the process of gathering genomic data by reading the nucleotides of a DNA molecule. This is done by a sequencer. There are different types of sequencing techniques (e.g Illumina (next generation sequencing), Oxford Nanopore, Pacbio (third generation sequencing), etc.) and different variants of sequencing (ATAC-seq, scRNA-seq, ChIP-seq, ...).

The simplified process of Illumina sequencing is as follows: Several DNA target sequences get treated with ultrasound in order to break them down into smaller fragments of a certain length with a certain margin of error (e.g. 200 bp +/- $x$ bp). These fragments are labeled, placed onto a flow cell, amplified, and sequenced. This generates an abundance of short reads, which often overlap with each other and contain mutations. The reads are then aligned to a Reference Genome via a Mapper like STAR in order to determine the original sequence. A Read Simulator is a tool that simulates this process of generating fragments and reads for given transcript sequences. In our case, we simulate a paired-end sequencing experiment, where we generate two reads for each fragment, one for each end of the fragment. With a Read Simulator, we are able to validate the results of Mappers since we know where our reads originated from. In other words, if a Mapper was to incorrectly map a read (generated by our Read Simulator) to the Reference Genome, we could detect this faulty behavior and see how far off the Mapper was to the actual coordinates. Like this, we could also detect entire regions inside the Reference Genome where the Mapper struggles in general (e.g. highly repetitive regions). A downside of this Read Simulator is that it assumes a normal distribution of fragment lengths and their starting position inside the transcript and a constant mutation rate across the entire read sequences, which in reality is not the case. However, it is still a valuable tool for testing the performance of other tools.

This figure was created using BioRender. For more information, visit BioRender.

Note

This figure is a simplification of the real underlying process (i.e. the mate read is generated by reading the reverse complement of the original fragment and maps to the opposite strand as first of pair etc.