Low coverage for bases near the end of target reference sequence

Question

[kspata]

Hi All,

I have viral samples sequenced using NextSeq PE 150 with over 90 million reads, I aligned these reads after trimming adapters and low-quality bases to a reference sequence (around 2750bp length). I noticed the bases at the end of the target sequence (2735bp - 2750bp) have very less coverage (equal to 0X or less than 10X) while the average per base coverage is 16700X, I used bowtie2 local alignment with default parameters.

Why am I getting low coverage near the end of target sequences? How can I improve coverage for this region?

Thanks in advance!!!

[h.mon]

Is this a RNA virus? Did you start sample prep from RNA?

[kspata]

I believe it is a DNA virus as sample prep was done using complete DNA purification kit.

[h.mon]

Is the drop in coverage sudden or smooth?

Never mind that, if it is just the last 15 bases, it is a sudden drop. In addition to genomax suggestions, it could also be an artifact. Are there similar viruses with the genome sequenced? Did you blast your whole virus sequence on NCBI?

[GenoMax]

Two things to consider:

• This may be a problem with aligners not being able to map reads (which may be much longer) to last 15 bp of the reference.
• You may have discarded reads with the last 15 bases during your trim/cleaning process.

[kspata]

Thank you for response. I did a two level trimming (trim_galore and sickle) which resulted in loss of reads aligning at the end of target sequence. I re-trimmed the raw reads with trim_galore only and got less than 10X coverage at ends of the target sequences. The coverage analysis still shows 0X coverage at base position 2749 -2750.

How can I align partial reads only to the ends of target reference sequence to get more than 0X coverage? Will this approach work and how can I use bowtie2 to do this?

[h.mon]

Is the genome of your virus circular or linear?

How are you looking at the coverage?

[GenoMax]

kspata : Have you considered the possibility that your viral strain has a small deletion at those two base pairs? So what you are seeing is real for your strain.