Working with a sequence with high similarity at nucleotide level but none at protein

[vpi]

Hi,

I have a sequence generated from an assembler which shows 95% identity to a nucleotide sequence when I BLAST or blastx it. But there is no similarity at protein level (blastp). Also, the translated sequence has multiple asterisks (*) , which means stop codons. If its at the end it means the end of the sequence, what does it mean when there are many of them?

I would like to do a homology model of the sequence. Any suggestions how to go ahead.

Any help is very much appreciated.

Thank you

[GenoMax]

That is strange. Is the identity to the protein/gene you are expecting? What kind of an experiment has the sequence come from (DNA/RNA seq)?

[PatrikD]

Are you sure you're looking at the correct reading frame for the protein sequence? If you start reading codons one base over, you'll just get random amino acids with a bunch of stop codons.

I would guess you either have a homopolymer sequencing error (reading one base too few or too many in a series of identical bases), a mis-called start of the ORF, or a pseudogene with an indel.

[gsgs]

can you post or upload an excerpt of the sequence ?

even if you "construct" a sequence with low protein similarity,
with 5% nonmatching nucleotides you can only get 15%
nonmatching amino acids

[vpi]

Thank you all for the replies!

Its a DNA sequence obtained from Miseq and using IDBA assembler. Just for the information, DNA sequence was translated using ExPASy, and I took the sequence from 5'3'Frame 1.

I tried to remove '*' and blastp, then I get few hits with very low identity to a bacterial genome, whereas, the nucleotide sequence shows high identity to a virus. Can I use such a sequence for homology modeling where stop codons are manually removed?

[gsgs]

can you also post the nucleotides-axcerpt

[gsgs]

I took the opportunity to update my virus database from genbank...
I'm not sure yet what virus it could be
probably none of the 3 big HIV,Influenza,Hepatitis

maybe the amino-acid sequence of that virus is not available ?

here are possible other reading frames:
[I use "}" for stop codons]

>test1
LVSKSVGSDCCYLLHCHSVMSKVN}CVQVLYSPRGSWAPHSNMFGFHVNADFCIYCDRIMLINYADSNG-
>test1 revert
QR}}IRWPSTSN}HRPYPRIWARPKPTWLCVVNI}KL}CPAGPHDWSG}TS}CYRSTCYHRLQAFGFTT-
>test1 revertf
NHCCPHS}}A}SGHNKCKNPRLHETQTC}SVVPTSP}GCKVLEHIS}LCS}RCGNGASSNNQNLRIWTP-
>test1 revert revertf
GCEPECLESVVAGRTVTL}RLP}PVVWTCRALKLLDVYNTEPRGFGTCPYSGVRAVSIRRRRPAYSLPL-
>test1
WCPNP}VLIVATCSIATAS}AKSTDVFKYFTALGARGHHTLTCLGFM}TRIFAFIVTGSCSLTMRTAMV
>test1 revert
NGSEYAGRLRRIDTARTPEYGHVPNPRGSVL}TSKSFSALQVHTTGQGRRHSVTVLPATTDSKHSGSQ-
>test1 revertf
TIAVRIVNEHDPVTINAKIRVYMKPKHVRVWCPRAPRAVKYLNTSVDFAHDAVAMEQVATIRTYGFGHQ
>test1 revert revertf
VVNPNAWSRW}QVER}HYDVYPDQSCGPAGH}SF}MFTTQSHVGLGRAHIRGYGRCQFDVDGQRIHYR-
>test1
GVQIRRF}LLLLAPLPQRHEQSQLMCSSTLQP}GLVGTTL}HVWVSCKRGFLHLL}PDHAH}LCGQQW-
>test1 revert
TVVNTLAVYVELTPPVPPNMGTSQTHVALCCKHLKALVPCRSTRLVRVDVIVLPFYLLPPTPSIRVHN-
>test1 revertf
PLLSA}LMSMIRSQ}MQKSAFT}NPNMLECGAHEPLGL}ST}THQLTLLMTLWQWSK}QQSEPTDLDT-
>test1 revert revertf
L}TRMLGVGGSR}NGNTMTSTLTSRVDLQGTKAFRCLQHRATWVWDVPIFGGTGGVNST}TASVFTTV-

[GenoMax]

@vpi: Have you blasted some of original reads to make sure the data is from the right organism. This would not be the first time there is a mixup of some kind (either in your hands or at the sequence provider)?

Possibility also exists that IDBA has produced a bad assembly.

[gsgs]

TTTACATGAAACCCAAACATGTTAGAGTGTGGTGCCCACGAGCCCC
,-3.nuc, JX514942, Human enterovirus C116


>test1 revertf
TTTACATGAAACCCAAACATGTTAGAGTGTGGTGCCCACGAGCCCC

>test1 revertf
YMKPKHVRVWCPRA-

inverse reading frame, starting at the 3rd nucleotide TAC...
this partial amino-acid sequence is also given in the genbank record

http://www.ncbi.nlm.nih.gov/nuccore/JX514942

[vpi]

Thank you all for the suggestions!

[kmcarr]

Thank you all for the replies!

Its a DNA sequence obtained from Miseq and using IDBA assembler. Just for the information, DNA sequence was translated using ExPASy, and I took the sequence from 5'3'Frame 1.

I tried to remove '*' and blastp, then I get few hits with very low identity to a bacterial genome, whereas, the nucleotide sequence shows high identity to a virus. Can I use such a sequence for homology modeling where stop codons are manually removed?

Why are you only translating in one frame? An assembled contig from raw reads is not expected to be in frame or even in the proper direction. You need to examine all 6 possible reading frames (3 forward & 3 reverse) for the correct coding sequence. Also, if there any misassemblies which result in "indels" in your contig this would cause the proper reading frame to shift in the middle of your coding sequence.

[vpi]

Thanks a lot @kmcarr....your suggestion was of great help. I looked at other frames and found that 3'5' frame 1 had a perfect sequence with no stop codons in between.