JigsawSeq

Comprehensive sequence profiling of gene variant libraries is important as it is broadly applicable for fine mapping studies in genotype-phenotype relations. JigsawSeq is a protocol to generate and evaluate a full-length gene variant library by combining codon-based molecular barcoding strategy. This program JigsawSeq is a package of perl scripts to perform de novo assembly with next-generation sequencing data from JigsawSeq experiments, and analyze the contigs.

Getting Started

JigsawSeq package is consisted of stand-alone perl scripts and the lastest version can be downloaded from github.

$ git clone https://github.com/jy2/JigsawSeq.git

If you would like to download the version (r3) used in the JigsawSeq paper published in Nature Communications, please clink here.

Then, you can run JigsawSeq on cloned directory without installation.

Requirements:

• perl 5.10 or higher

• bwa 0.79a

• samtools 0.1.19

cf. bwa and samtools are included in this package.

Simply, JigsawSeq can be excuted like:

$ ./Jigsaw.pl -F Ex1_F.fastq -R Ex1_R.fastq -V pBR322_vector.fasta -L 816 -O Ex1

The example fastq files (Ex1_F.fastq, Ex1_R.fastq) can be downloaded in here; Ex1_F.fastq.gz and Ex1_R.fastq.gz.

With default parameters, JigsawSeq will use ~20Gb of memory for analyzing a 1.5Gb size of fastq file (~5,000,000 reads for KanR library). But, the memory usage depends on several factors; amount of reads, length of contigs, mutation rate, and sequencing quality. Therefore, the exact memory usage will be varied at specific situations. We highly recommend to use Hiseq data instead of MiSeq or NextSeq data, since sequencing error rates of MiSeq/NextSeq data are generally higher than one of HiSeq data, and much higher memory space is required when using other machines. Note that you can reduce the memory usage by lowering --bin_size (default: 15,000,000) when running Jigsaw.pl.

The final analyzed results will be stored at ex1.contigs.result.This result file is a tab-delimited text file with columns; contig name, length of contig, depth of contig, initial seed, terminal seed, and contig sequence. The numbers on 4th and 5th column are matched to the seed sqeuences at the header.

Workflow of JigsawSeq

Commend-line options

Jigsaw.pl

• Description: main module to preprocessing raw fastq files and construct contigs.

• Usage: Jigsaw.pl [OPTIONS] -F [input:fastq] -R [input:fastq] -V [input:fasta] -L [input:integer] -O [output:prefix]

  • Parameters: Among the parameters below, -F, -R, -V, -O, -L are required parameters

    • -F, --forward Input filename of forward reads (fastq format)

    • -R, --reverse Input filename of reverse reads (fastq format)

    • -V, --vector Input filename of vector sequence (fasta format). The representation of vector sequence must be followed; 3'side of the insertion site must be the first position of vector sequence and 5'-end of the insertion site must be the last position.

    • -O, --output Prefix of temporary and final output files

    • -L, --length Expected length of contig will determine the maximum number of steps on exploring graph.

    • -k, --kmer Length of k-mer. Default = 120

    • -s, --step Step size (1-, 2-, or 3-mer) for exploring the graph. Default = 3

    • -m, --min_depth Minimum depth of nodes and edges. The nodes and edges with lower depths than min_depth will be excluded. Default = 2.

    • --cut_edge Cutoff ratio for edges. To save the exploring time and memory usage, and to avoid false positive contigs by errorous edges the edges will be discarded if the depth of edges divided by the maximum depth of edges linked to the same node is lower than 1/cut_edge. Default = 150.

    • --min_seed Minumum depth of seeds. The seeds with lower depth than min_seed will be ignored. Default = 100.

    • --cut_seed Cutoff ratio for seeds. To avoid false positive contigs by errorous intitial/terminal seeds, the seeds will be neglected if the depth of seeds divided by the maximum depth of seeds is lower than 1/cut_seed. Default = 100.

    • -e, --exclude File name of problematic sequences (fasta format). Any nodes having these sequences will be excluded. Default = exclusion.fa

    • --read_length Read length of raw reads. Default = 150

    • -t, --thread Number of threads. Only for detecting seeds by using bwa. Default = 3.

    • -b, --bin_size Number of reads in a bin. Lowering --bin_size will reduce the memory usage, but increase the running time. Default = 15000000.

    • -C, --cut_CV Cutoff for coefficient of variation. This will be used only if the contigs were re-aligned by raw data. (-a option) Default = 0.2163.

    • -a, --realign Allow to realign raw reads to contigs. This mode will reduce false positive contigs but decrease sensitivity.

    • -v, --verbose Verbose-mode On. Temporary files will not be deleted.

Information

• Version: r3

• Last modified: Apr-22-2015

• Contact: Jung-Ki Yoon M.D. (dr.jkyoon at gmail.com)

• License: JigsawSeq is released under a CC BY-NC-SA license.