3
$\begingroup$

I have an assembly of an algae and want to find if it has anything analogous to endogenous retroviruses' subsequences in its DNA.

My guess is I should manually blastn all subsequences until I find (or not something) that could have come from a virus.

Is there a better way to do this?

$\endgroup$
2
2
$\begingroup$

Your guess sounds reasonable to me but you should consider using KRAKEN2. You can create your custom database (you would have done it with makeblastdb anyway) and kraken will check k-mers against the database to find homologous sequences.

$\endgroup$
4
  • $\begingroup$ I don't feel Kraken is good for this task. Kraken taxonomically classifies each sequence, while the signature of integrated viral DNA will be a single sequence with both host and virus parts $\endgroup$ Aug 17 '20 at 14:56
  • $\begingroup$ @Chris_Rands. Preparing your custom dataset with only viral genome and with the right k-length for the kmer analysis is the only method I know of existance today. You can use manually blastn (honestly, I do this myself if it's not for publication) or let the job to a well developed (and fast) software out-there. Please, let me know if you think of a better or potentially better solution! $\endgroup$
    – aerijman
    Aug 17 '20 at 15:08
  • $\begingroup$ Your suggestion might be fine for detecting viruses in general, but not integrated viruses specifically- this requires a fundamentally different approach because you are looking for chimeric reads $\endgroup$ Aug 17 '20 at 15:25
  • $\begingroup$ @Chris_Rands chimeric reads could be found using shorter kmers (at the expense of dropping specificity). Do you know of any better or alternative solution? $\endgroup$
    – aerijman
    Aug 17 '20 at 15:29
-2
$\begingroup$

Searching directly for a reverse transcriptase may find recently incorporated retroviruses.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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