For full documentation of the 2bRAD de novo pipeline see the github page
The pipeline is very similar to that performed by Stacks (Catchen et al. 2011):
De novo locus generation
#navigate to the directory #look at starting trimmed fastq files ls *.trim sampleA.trim sampleB.trim sampleC.trim #run uniquerOne.pl #(this is analogous to making 'stacks' in STACKS (Fig1A Catchen et al. (2011)) #finds the unique RAD tags from each fastq uniquerOne.pl sampleA.trim > sampleA.trim.uni uniquerOne.pl sampleB.trim > sampleB.trim.uni uniquerOne.pl sampleC.trim > sampleC.trim.uni # merging uniqued files #(Fig1B Catchen et al. (2011)) mergeUniq.pl uni minDP=2 >mydataMerged.uniq #generates a merged set of unique tags: mergedUniqTags.fasta # clustering allowing for up to 3 mismatches (-c 0.91); the most abundant sequence becomes reference #This is equivalent to calling loci (Fig1C-D Catchen et al. (2011)) module load cd-hit cd-hit-est -i mergedUniqTags.fasta -o cdh_alltags.fas -aL 1 -aS 1 -g 1 -c 0.91 -M 0 -T 0 #now we have called de novo loci based on the tags #assemble them into an artificial reference for re-mapping and genotyping concatFasta.pl fasta=cdh_alltags.fas num=8 #index the artificial reference with bowtie module load bowtie bowtie2-build cdh_alltags_cc.fasta cdh_alltags_cc.fasta #now map the reads back to the artificial reference bowtie2 --no-unal -x cdh_alltags_cc.fasta -U sampleC.trim -S sampleC.trim.bt2.sam bowtie2 --no-unal -x cdh_alltags_cc.fasta -U sampleB.trim -S sampleB.trim.bt2.sam bowtie2 --no-unal -x cdh_alltags_cc.fasta -U sampleA.trim -S sampleA.trim.bt2.sam #The alignment files can now be used for whichever genotyping method you prefer