Especially choice of aligner, different cut offs, whether to do duplicate removal cause RNA seq libraries are enzyme cut libraries, minimum depth of coverage for variant calling, rare allele minimum frequency etc.
I'm using Tuxedo pipeline (http://tophat.cbcb.umd.edu/). It uses Bowtie2 aligner and is the best one for me. To remove duplicate I use the 'rmdup' tool from samtools in the bam file generated on tophat mapping. I don't think that depth coverage in RNA-seq experiments is really necessary. You shoud worry about the biological and technical replicates, it will make a big difference in your analysis. Samtools have a lot of cool inplementations, you'll be surprised what it can do.
So, what tools you try to mean in Samtools set. Why to do remove duplicates here.? And why not depth of coverage has a cut off here?
Duplicates affect mainly the differential expression analysis. Duplicates are atificially generated during Illumina sequencing, it's not represent your "real" reads so you have to eliminate this bias. If I'm not wrong, Samtools make SNP call. For this purpose I believe ate least 50X is necessary but the coverage in transcritome is a little tricky because the differential coverage between genes. I never see a paper saying a explicit cutoff value, you can look for what people are using for your organism.
Rohan, can you please share the cutoffs... I have a good pipeline, but, I am not sure about the cutoffs that would be good.
Are you talking about dynamic cut off... mine data set is illumina pairend, human, 2 technical replicate per sample large biological replicate (cultured cell, before treatment after treatment set up). Present cut off- at least 10 reads to call any variant, atleast 5% to call rare variant, 15% to call novel variant and known deletion, 20% to call deletion. What are your suggessions on these?
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Bioinformatics
- Bioinformatic Tools