After aligning paired-end reads to a reference genome, I am getting low percentage:

34636166 + 0 in total (QC-passed reads + QC-failed reads)
34346076 + 0 primary
0 + 0 secondary
290090 + 0 supplementary
0 + 0 duplicates
0 + 0 primary duplicates
5706057 + 0 mapped (16.47% : N/A)
5415967 + 0 primary mapped (15.77% : N/A)
34346076 + 0 paired in sequencing
17173038 + 0 read1
17173038 + 0 read2
3807784 + 0 properly paired (11.09% : N/A)
3996748 + 0 with itself and mate mapped
1419219 + 0 singletons (4.13% : N/A)
177284 + 0 with mate mapped to a different chr
65388 + 0 with mate mapped to a different chr (mapQ>=5)

What has to be done to increase the mapping percentage?

I'm attaching the fastqc report before trimming and after trimming too first- R1 without trimming second- R2 without trimming

third- forward paired fourth- reverse paried

rest are overrepresentations

After using trimmomatic the quality is getting more bad and even the mapping is of low percentage

Mash screen result in the last image

enter image description here

enter image description here

enter image description here enter image description here

enter image description here

enter image description here

  • $\begingroup$ My initial guess would be that you have a low quality reference genome, or you mixed up the reference genome, or you have significant contamination of your library, or that your sequencing run is very low quality. I'd suggest running fastqc and posting the results here, posting some statistics about your genome, and generally explaining where your reads are coming from and what your reference is. $\endgroup$ Aug 4 at 17:29
  • $\begingroup$ @MaximilianPress Thanks for your reply . I have the data of Parantica aglea I done fastqc that report was showing probelm of gc content and base quality even after trimming the sequence reads with trimmomatic it got more worse $\endgroup$
    – Moon
    Aug 7 at 5:52
  • $\begingroup$ I am unclear if this solved your problem, possibly consider posting the fastqc outputs if you want more help. Seems like the reads might just be bad but hard to evaluate from what you say. $\endgroup$ Aug 7 at 16:30
  • 1
    $\begingroup$ The reads honestly look fine to me from a sequence quality standpoint. FASTQC thinks that the basecalls are good based on the green light. Your first few nucleotides are a little weird compositionally but that's probably due to UMIs or barcodes in your library, without seeing more information. A read mapper would just soft-clip those. We are not seeing e.g. an excess of Gs that you would expect from an instrument failure. You are not showing the overrepresented sequences plot which is probably most likely to show any systemic issues, but none of the plots that you are posting look bad. $\endgroup$ Aug 9 at 21:21
  • 1
    $\begingroup$ I'd suggest going through the hypotheses in my first comment. The reads look at least sort of ok (though you should check the overrepresented sequences tab of Fastqc). So you could look at 1) contamination, try using mash screen 2) wrong/bad reference genome (you will know better how to do this, maybe mash screen the reference genome against the screen database?) 3) sample swaps of your library. I don't really have more to say than that. $\endgroup$ Aug 10 at 19:30

1 Answer 1


Just answering, following extensive back and forth in the comments.

I think that your reads are fine:

  1. FASTQC doesn't see anything too weird. You see some noise at the first few basecalls but I am not too alarmed about that.
  2. FASTQC finds overrepresented only a telomeric repeat and its reverse complement (TTAGG). That is something you'd expect to find in WGS reads.
  3. Mash screen finds that the only match, with a small but non-trivial number of k-mer matches, is the monarch butterfly Danaus plexippus, e.g. another lepidoptera. So that makes sense.

Therefore, the issue is likely that your reference genome has issues or has been swapped out with something else.

Where did your reference assembly come from?

  • Did it get mixed up somehow?
  • What does it look like when you run QUAST?
  • What does it look like when you run mash dist of your assembly against the same mash database? Is it maybe a bunch of endosymbionts or other contaminants?
  • What happens if you map your reads against a high-quality lepidopteran assembly like the monarch butterfly or this?
  • What happens if you map your assembly against a high-quality lepidopteran assembly? Is it syntenic?
  • $\begingroup$ I got the reference genome from ncbi database. I tried out mapping with another reference genome of same subfamily which is Danaus chrysippus and the result are same and there are no reference genome available apart from this so I'm thinking might be there is some genetic variation between the parantica aglea and Danaus plexippus . So, de novo assembly will the solution for it ? or is there any another solutions which I can look for ? $\endgroup$
    – Moon
    Aug 24 at 6:38
  • $\begingroup$ @Moon It would be helpful if you could specify the reference assembly initially in the future. I still don't know which reference assembly you are working from, but I am guessing that it is not in the same genus as your reads. Yes, you can expect significant variation even between fairly closely related organisms- 10-20% is not crazy unless they are pretty close. Yes, I would do a de novo assembly. Even with just Illumina you are likely to do ok compared to these other references. $\endgroup$ Aug 24 at 21:23
  • $\begingroup$ I have the sample sequencing reads of Parantica aglea and the reference genome I'm using is Danaus plexippus which comes from the same subfamily I was trying to do Reference-guided genome sequence assembly pipeline so I initially started with fastqc and then trimming and here I'm on bwa mem where there I'm getting low percentage even after changing the reference genome and software tools the story is still same as you said might be the issue with reference genome so de novo assembly can be done I thought. Might this will clear what I'm working on $\endgroup$
    – Moon
    Aug 28 at 8:54
  • 1
    $\begingroup$ @Moon I would suggest not doing a reference guided assembly. I would instead do a de novo assembly. I expect that the issue is that your genome and the Danaus genome are substantially different in structure and you will not be able to find anything meaningful by reference guided assembly, and may be actually misled. After you perform the de novo assembly you can still try to scaffold the resulting contigs using the Danaus reference using a tool like RagTag. $\endgroup$ Sep 1 at 21:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.