Using FastQC, I have checked the over all quality of the file. The output file contains much count of over represented sequences. I simply performed blast on these sequences, and surprisingly they have given to hit some gene which are really important for the analysis. Results are attached
So is it necessary to remove over represented sequences in the NGS data?
I concurr. It is always safer to include the whole data set. But depending on what analysis you want to perform You can work on arepresentative set.
if we include whole data set, it may result in poor read mapping. Isn't it?
It's good to remove adapters, perform quality checks and you can even look computationally within the sequences for overrepresented data. There should be no mechanical error or noise present while handling samples. Look for certain thresholds else you may miss some important information.Considering whole dataset may not affect read mapping but may result in duplication of information.
Its better to remove over represented sequences, adapters and low quality bases at the sequence read ends which improved the assembly significantly. So prefer to do trimming.
"Trimming" and "removing" technically the output will be the same. Is'nt it?
- Badges
- Science method