Entering edit mode
3 months ago
I'm analyzing a dataset of GEO containing RNA seq data of 23 samples using galaxy server. I've mapped reads of the data through STAR but there is a problem, 4 samples were aligned less than 70% (Even 2of theme were aligned 40%). How should i solve this issue ?
What is the problem that needs to be solved?
Maybe the dataset has lots of adapter diners, or rRNAs, or other contaminants, etc.
If you want to solve it, take the unaligned reads and manually inspect them (e.g. run them through BLAST).
Thank you. I have rechecked the alignment's results. The main issue is mapping to multiple loci rather than unmapping. Is the cause contamination? Isn't it due to the design of experiments and the analysis of mRNA expression?
At Edinburgh Genomics we take RNA-seq data QC seriously. We assemble the unmapped reads into putative transcripts, then use 'blobtools' and 'diamond' to assign a likely taxon to umapped reads and assess for potential contamination from another species.
When preparing short read RNA-seq libraries it can be hard to control the fragment size, so reads containing some adapter are common. Checking the 'insert size' (for instance using picard tools) is useful.
If you are studying a non-model organism, be aware that in complete reference genome, or a reference genome not so closely related to your sequenced individuals, can also result in a low mapping rate.
I've rechecked my data, there were a few unmapped reads the main issue was mapping to multiple loci. Does this issue affect my study ? What should I do in this case ? (My study aims to analysis the expression of genes through RNA seq)
Couple explanations eg low quality reads, number of reads, adapters, etc. you will need to do some basic quality control using a tool such as FastQC
Thank you. I already used FASTQC for quality assessment. There were no errors and quality scores were good. Though I think this issue is due to contamination.