relation between Illumina sequencing primer and viral sequences

Question

Dealing whith a problematic sequencing run I found this over-represented sequence:

GGAAGAGCACACGTCTGAACTCCAGTCACTAGCTTATCTCGTATGGCGTCTTCTGCTTG

It is clearly relate to the Illumina sequencing primer as shown by this read structure description (note the bold residues above):

I searched the sequence in the nucleotide NR ncbi database, using blastn (standard parameters) and I found this as a first hit:

Severe acute respiratory syndrome coronavirus 2 genome assembly, chromosome: 1
Sequence ID: OV196492.1Length: 30243Number of Matches: 2
Range 1: 73 to 133GenBankGraphicsNext MatchPrevious Match
Alignment statistics for match #1
Score   Expect  Identities  Gaps    Strand
89.8 bits(48)   2e-14   57/61(93%)  2/61(3%)    Plus/Minus
Query  1    GGAAGAGCACACGTCTGAACTCCAGTCACTAG--CTTATCTCGTATGGCGTCTTCTGCTT  58
            ||||||||||||||||||||||||||||| ||  ||||||||||||| ||||||||||||
Sbjct  133  GGAAGAGCACACGTCTGAACTCCAGTCACCAGTGCTTATCTCGTATGCCGTCTTCTGCTT  74

Query  59   G  59
            |
Sbjct  73   G  73


Range 2: 188 to 248GenBankGraphicsNext MatchPrevious MatchFirst Match
Alignment statistics for match #2
Score   Expect  Identities  Gaps    Strand
89.8 bits(48)   2e-14   57/61(93%)  2/61(3%)    Plus/Minus
Query  1    GGAAGAGCACACGTCTGAACTCCAGTCACTAG--CTTATCTCGTATGGCGTCTTCTGCTT  58
            ||||||||||||||||||||||||||||| ||  ||||||||||||| ||||||||||||
Sbjct  248  GGAAGAGCACACGTCTGAACTCCAGTCACCAGTGCTTATCTCGTATGCCGTCTTCTGCTT  189

Query  59   G  59
            |
Sbjct  188  G  188

There are also sequences of other virues that match very well.

Do you know what is the relation between Illumina adapters/primers and viral seqiuences?

Answer 1

You'll unfortunately find adapter sequences contaminating a lot of entries in Genbank. In short, the sequence is most likely either an unassembled fragment where the input into the assembly step hadn't had adapters trimmed or (somewhat more rarely) the adapter sequence was incorporated into an assembled contig and then uploaded. You'll see a lot of lower quality sequences like this, especially from massive viral screens.

Be aware that there are other quality issues facing particularly virus sequences such as extremely high NNN stretches (cf. the difficulties faced in creating the C-RVDB viral database once there were hundreds of thousands of apparently poor-quality coronavirus sequences automatically added to Genbank).

Answer 2

SARS-CoV-2 NCBI no. OV196492.1

Summary The relationship sought is simply the genome position because this is the UTR (UnTranslated Region).

The linker is present starting at position 74 and is also present at position of 189 of the SARS-CoV-2 genome from 5' to 3' (I checked). This is where position 1 is not necessarily position 1 of the genome, but the start of the authors genome. So it's the UTR. I did suspect this was the location.

There are three reasons for the drop in sequence quality at the UTR:

1. overall quality, really the read depth, of short-read sequence data is reduced at the terminal regions of viruses per se;
2. there is extensive secondary structure in the UTR regions which is critical to its function, thus the folding will affect short-read sequencing approaches, e.g. primer binding will be more difficult because these sequences will preferentially self-anneal (preventing the primer from binding);
3. QC annotation from the authors will focus on the coding regions, the UTR is outside the coding regions and therefore can get missed.

In this particular case point 3 is likely the underlying reason. The authors QC ignored the UTRs (the 3' UTR is also problematic). If a linker is caught in the coding domains the translation will flag it immediately ... obviously the UTR isn't translated. Thats simply why. For virologists, this isn't cool, but for clinicians they don't see its importance.

Overall, this is simply an issue of ad hoc QC because the NCBI SARS-CoV-2 database is extremely powerful and excellent for custom analysis. To "throw in the towel" on the power and flexibility of NCBI's SARS-CoV-2 database for what is essentially some personal QC coding isn't worth the trade-off in my personal opinion. It's a powerful portal, which enables an analysts to focus on downstream issues.