Depends what you want to do and what kind of sequencing data you are talking about. If it is just a sanger small fragment sequencing of 300-1000bp i would send it again because it is cheap.

If it is for example illumina hiseq NGS then you can trim low quality data at the 3' and 5' ends where they usually show up and continue with the higher quality data.

Cheers

Dear Bassiaka Ouattara,

You can redo the sample and give it for sequencing. Or else try AB biosystems for sequencing, it will give a good quality of sequence. 

Best of Luck

You may communicate with the support team of the company and discuss possible reasons for the low quality. Also, if NGS, the company may have stated a typical output and depth of a sample, e.g. 80% of sequences at 20x depth at quality level XX. This can be a reference for your discussion. If you have access to files of mapped sequences you can browse (IGV, GenomeBrowse) along chromosomes to find any regions with poor quality and see if it is end-of-reads related.

Best

/Hans

First of all: was your DNA good? Did you purified your samples? Usually in my group we run an agarose gel to check amplification and if we have too much primer-dimer we purify the samples with Exosap (for instance) and only then we send it to sequence. If the remaining primer-dimer has as much intensity as my main band I know the sequencing quality will be too low. If none of this applies I would ask to repeat at least a few of them and then compare with the first run. If quality improves then it was something related to the sequencing run. If everything stays the same it might be your samples that were not good in the beginning.

I hope this helps.

My work was an assesment of genetic divereity using internal transcribed spacer sequences of nuclear ribosomal DNA. DNA sequencing was done through NGS technologies (illumina). It is about 800-1000 kb and a total of 96 individuals. Unfortunately I don't have opportunity to run again the sequencing. I thought it is possible to treat low quality of sequencing without running again. According to the different answers, that seems impossible. I heard that when the QC values are high than 22 the data are reliable. Are there? Thanks to all

Thank you Artur a lot for the interest

I'm not expert in bioinformatic but I can do some simple analysis. I ordered a pair-end sequencing. For each individual, I used forward and reverse to make contig assembly through DNAbaser. For almost of my sample, I trimed ends until there was 40% good bases in a 13 bases window ( QV of each base is about 15). This is the max quality I had. I read that the quality I have is not good. They said I must have at least 60% of good bases in a 16 bases window (each base must have a QV higher than 20 in order to be considered good) but I don't have this. Most of my base has a QV low than 20. This is my problem.

For private message, you can join me at [email protected]