PAR CLIP analysis

[panosk]

Hello everyone. I am new here so please let me know if my post does not fit forum's appropriate posting policies.

I am a bioinformatics msc student and I am trying to analyze a dataset with PAR CLIP reads (http://www.ncbi.nlm.nih.gov/geo/quer...acc=GSM1020024). First thing is alignment.

I have used FastX Artifacts Filter, Trimmomatic, FastX Trimmer and Cutadapt to prepare the data set for alignment.

I have used TopHat 2 aligner without success (5% alignment) following the way mentioned here : http://www.ncbi.nlm.nih.gov/geo/quer...?acc=GSM994709

Any clues why am I getting such a low percentage? Is it because of the T to C conversions by the presence of 4-thiouridine? Am I missing something?

I am about to try with BWA (BWA-Backtrack) and GSNAP (If I solve an illegal instruction error I get) too.

Thanks in advance.

[Orr Shomroni]

I would suggest you to run FastQC for the pre-processing files in the different steps, as this will show you the sequencing quality of your data (just useful to see none of the steps completely messes up the reads). As for the alignment, personally I ran Bowtie2 with 1 mismatch on some PAR-CLIP data I have, and I get very low alignment percentages (a bit more than 5%, but less than 50%). I am considering to try STAR, since you are dealing here with sequences based on RNA, and STAR is considered to be one of the best alignment tools for RNA-seq.

In addition, I don't know how much you've read about this technique, but I would REALLY recommend you starting with the first PAR-CLIP paper by Hafner et. al.

[panosk]

Thanks for your reply. I have used FastQC and I have also read (also) that paper. I have tried many approaches. STAR needs 30GB RAM, something that my availability of resources cannot support. I have tried with bowtie,tophat and bwa(backtrack) without a big success but time is pressing me so I carried on with what i have.