prep

Mapping instructions to transcriptome and/or genome

• This instruction is used to map the RNA structure profile data to transcriptome and/or genome.
• Finally the file is converted to BigWig format.

Step 0: Requirements

Tools	Link
bowtie2	http://sourceforge.net/projects/bowtie-bio/files/bowtie2/2.2.5/
samtools	http://samtools.sourceforge.net/
ngslib	https://github.com/tsznxx/ngslib/wiki
rsq	https://github.com/tsznxx/rsq/wiki

Step 1: Build index using Bowtie2-build

• Remove sequences like chr*_XXXXX from hg19_all.fa

# Filter out chrom names containing "\_"
>awk '{if ($1 ~/>/) {if ($1 ~/\_/) choose=0; else choose=1} if(choose == 1) print}' hg19_all.fa >hg19.fa
# Filter out chromosome names with "_"
>awk '{if($3 !~/\_/) print}' hg19_RefSeq_all.genepred >hg19_RefSeq.genepred

• Remove non-polyA RNAs as many as you can from hg19_RefSeq_all.genepred

# Custom designed script to remove small noncoding RNAs (starts with NR_ and shorter than 200 nt)
> python filterout_sncrna.py hg19_RefSeq.genepred >hg19_RefSeq_rmsncRNA.genepred

• Remove genes on 'chr*_XXXX' from hg19_RefSeq_rmsncRNA.genepred
• Relabel duplicate transcript names. Duplicate names can be shown as NR_110999:2/4, which means this is the 2nd of the 4 replicates.

>wCleanTranscriptome.py hg19_RefSeq_rmsncRNA.genepred > hg19_RefSeq_clean.genepred

• Build Bowtie2 index

>wBedToFasta.py -i hg19_RefSeq_clean.genepred -g hg19.fa -f genepred -o hg19_RefSeq_clean.fa
>bowtie2-build hg19.fa,hg19_RefSeq_Relabel.fa hg19_merge

Note: When wBedToFasta.py runs, an index file is built for hg19.fa as hg19.fa.fai. This file contains chrom names and sizes in the first two columns, and can be used in the following BAM->BigWig step.
An alternative way to generate a chromosome sizes file is using UCSC executable faSize:

>faSize -detailed hg19.fa >hg19.sizes

Step 2: Mapping and sorting

• Using bowtie2 default reporting strategy, so that at most one alignment is reported for each read.
• Parameters in red are set specifically for the data we are using. Users can tune alignment parameters according to their own data quality.

>bowtie2 -3 50 -p 10 -x hg19_merge -1 GM12878_S1_1.fastq -2 GM12878_S1_2.fastq|samtools view -Sbh - |samtools sort - GM12878_S1
>samtools index GM12878_S1.bam

Step 3: RNA structure profile extraction (BAM->BigWig)

• This step is technology dependent. Here we take PARS data as an example.
• The upstream -1 position is the recognition site.
• Multiple BAM files are accepted. BigWig files will be created for each BAM file.

>wProfileExtraction.py --input GM12878_S1.bam GM12878_V1.bam --anno hg19_RefSeq_clean.genepred --chrsizes hg19.fa.fai --shift -1