We sent for sequencing several bacterial strains in batch to an external provider.
When we got our results back and started to analyze them, we found out that several of our samples were not what they were supposed to be. We sent toxic and non toxic strains and they were mixed up; also some strains that were supposed to be type A were type B in fact.
These sequencing results were not concordant at all with the biological results we got before.
We started to think it would be possible some sample names were exchanged, whether at our lab or at the sequencing facility.
We got the confirmation by checking the presence of specific markers genes in the sequences we got and compare it with the biological results we had.
Sequence of strain A1 showed to contain a proteolytic gene ptr1; we check by PCR the DNA of strain A1 for ptr1 PCR marker and it ended up negative (of course running positive and negative control at the same time). This was the confirmation that our sequencing results were wrongly named.
Now, the tricky things is to identify the mistake and to be able to correlate the sequences with the real strain name.
To investigate our wrong named sequencing results, we had the idea to check the Crispr content. Fortunately, our sequences contained some pretty different crispr spacers sequences.
Crispr typing is used to type Salmonella (Genetic analysis and CRISPR typing of Salmonella entericaserovar Enteritidis from different sources revealed potential transmission from poultry and pig to human, Li et al, 2018).
Therefore, based on spacers identification we designed specific sybrgreen PCR primers to amplify between 2 spacers (size of the spacer 36, DR 30, total amplification length less than 100bp). Of course, the spacer chosen were specific of the sequence investigated and did not match with any of the other sequences.
Then, we performed the PCR on the DNA we had at the lab. The results confirmed that the sequences we got were not related to the names of our DNA. PCR primers A amplified DNA of strain B, PCR primers B amplified DNA of strain C, etc. It turned out that one DNA was omitted while filling up the 96 well plate sent to the sequencing facility, shifting up the order of the names.
Of course we were lucky that the DNA sent to sequencing were different enough to perform Crispr typing. If it would have been clonal strains, we would had gone for SNPs identification.
Also, it would have been possible to sent back the DNA for sequencing but then it would had been much more expensive.
And it was kind of fun to play Sherlock Holmes with these sequencing results, trying to find out who was who, although it cost us a little more money and extra time to figured it out ;p.
this link is useful
https://www.nucleics.com/DNA_sequencing.../DNA-sequencing-mixe...
regards
HI Cedric
the best way to avoid those mistakes would be for you first to carefully annotate the samples and strictly follow all points of the procedure. a second point that could help is to prepare yourself the library and therefore barcode all samples before pooling.
hope this helps
fred
First, thanks for your feedback.
Hassan, your link is not working.
Frederic, I totally agree with you that carefully annotated samples are the basics of good laboratory practice to avoid such mistakes.
Using multichannel pipette is also helping when filling up a 96 well plate manually.
We did not prepared the library ourselves in this case.
For the record, this was posted to share with the community that it is possible to use Crispr typing to investigate mixed up sequencing results.
The problem was solved :).
- Badges
- Science method