1
$\begingroup$

I am assembling a transcriptome obtained from the Internet. The transcriptome was extracted from a human cancer tissue that had been previously grafted into a mouse. I have detected that many assembled RNA sequences do not match precisely to their reference but have additional sequences at one of their tails containing bacterial DNA. It's not that there are reads pertaining to that bacterium (always the same one) but that the RNA sequences of the cell have incorporated fragments of bacterial sequences (the reads corroborate these insertions with high coverage), therefore I can't just filter out bacterial reads as there are not reads exclusively pertaining to bacteria but hybrids. My question is how can it be possible? What possible mechanisms could have brought these bacterial fragments to the RNA of the cell. I have read in some papers that there is speculation that bacteria could be involved in the genesis of cancer cells. That would be an explanation if it wasn't because there are also mouse RNA for which the same circumstance applies: they also have bacterial sequences at the tails. I have also read in papers that there is speculation with the possibility that cancer cells are more susceptible to get DNA from bacteria transferred to their DNA. I am speculating that is the reason in this case (and somehow that weakness was transferred to contiguous mouse cells) but I would like to know if there is another possibility I am missing. Thanks.

$\endgroup$
6
  • $\begingroup$ Sorry, I mean there are bacterial sequences at the tail of the contigs, not only the reads. $\endgroup$
    – juanjo75es
    Jul 10 '20 at 18:55
  • $\begingroup$ Sounds like an opportunistic infection, which bacteria? Just blast it, anything pathogenic, or just coliforms ... if its mixed bits of genome from the same species - thats your answer. The mouse is very sick these things happen $\endgroup$
    – M__
    Jul 10 '20 at 21:51
  • $\begingroup$ The bacteria is Ralstonia solanacearum. It's mixed with mouse RNA and with human cancer RNA. there is no mix between mouse and human RNA. I'm trying to understand the mechanism by which the bacterial sequence ended up in the cancer and mouse RNA. $\endgroup$
    – juanjo75es
    Jul 11 '20 at 21:46
  • $\begingroup$ Usually it is like @swbarnes2 mentioned, adapter sequence. Can you check the fragment length of your aligned sequence? If they are short, then i think you can do an adaptor trimming first $\endgroup$
    – StupidWolf
    Jul 13 '20 at 18:39
  • $\begingroup$ @StupidWolf, the length of the reads is 140 bps and the contig is around 1400 bps for a transcript that should be 1280 bps long (GAPDH). That's only one of the transcripts. Not sure if you asked that. If I remove the adapters there are still more than 100 bps of bacterial sequence. $\endgroup$
    – juanjo75es
    Jul 13 '20 at 19:45
1
$\begingroup$

From time to time these situations are encountered, notably when PCR is being deployed in rare DNA samples.

In this instance the wiki states,

Ralstonia solanacearum is an aerobic non-spore-forming, Gram-negative, plant pathogenic bacterium.".

I'd leave it. The simple explanation is there has been a bad mix up. The best explanation is the people loading the Illumina machine, designating the adaptors etc ... In this scenario it has nothing to do with bacterial genome integration, plant bacteria don't touch mice and aerobic bacteria wouldn't grow in cancer cells.

The second issue is what is a soil bacteria doing in a mouse cancer biological facility? For the staff this wouldn't be good news.

The singular biological possibility is its a Pseudomonas and some species are serious opportunistic pathogens of immune-weakened humans, but there are a lot of Pseudomonas species and this one doesn't ring any bells, BUT its mice so who knows? One species in particular is associated with hospital outbreaks, but a biomedical unit that isn't quite the same thing.


The pathogen in humans is called Pseudomonas aeruginosa. If hospitals get this infection it is BAD NEWS, for example burns victims are really susceptible and without intervention it kills them. If you have evidence that the bacterial equivalent in mice is the bacterial pathogen you have found that is a substantial step. Biological Safety Facilities (BSFs) DEFINITELY get infected. They hate it when this happens and can be very picky for fear of a contamination, but it certainly happens. If you have evidence of a major contaminent pathogen in the long run it would be of HUGE help to BSFs globally. No-one really bothers with the disease pathology of mice - because their mice and spend their entire lives going 'squeek' - but when they are models for cancer their status changes beyond recognition. So I'd erase my initial thoughts and suggest, what you have found is publishable and the application is BSF contamination - don't say that though say this could be the mouse equivalent of P. aeruginosa. Its certainly a publication and goodluck.

$\endgroup$
6
  • $\begingroup$ Thanks Michael. I guess your answer is correct but I am still trying to understand how a bad mix up can result in that data. $\endgroup$
    – juanjo75es
    Jul 13 '20 at 13:09
  • $\begingroup$ Hi @juanjo75es the best way to thank me is to upvote or even accept :-) There is a possibility that the BSF (Biological Safety Facility) has been contaminated, it certainly happens and what you are looking at is the mouse equivalent of Pseudomonas aeruginosa (hospitals hate this disease).. If you can suggest this in publication that is the mouse equivalent of P. aeruginosa in humans for mice that is a major step forward. The BSFs of this world will not want to hear this news ... but who knows right? I'll update the post to incorporate this info $\endgroup$
    – M__
    Jul 13 '20 at 13:14
  • 1
    $\begingroup$ Thanks again, Michael! I will accept an answer once I really understand what's going on. Sorry but I am just developing software and have little experience in biology so I may be a bit slow. I see now @swbarnes2 answer saying that this is not bacterial contamination (if I understand well) so ... I am still confused. $\endgroup$
    – juanjo75es
    Jul 13 '20 at 19:07
  • $\begingroup$ Thanks @juanjo75es $\endgroup$
    – M__
    Jul 13 '20 at 19:21
  • $\begingroup$ Looks like the sequence is just an adapter that was incorporated into the Ralstonia assembly (and other ones) and that's why it matched bacterial genome. I guess I will have to accept another answer despite your kind help which is mysteriously uncommon in bioinformatics forums. I apreciate your interest to help. $\endgroup$
    – juanjo75es
    Jul 14 '20 at 10:33
1
$\begingroup$

What exactly is the sequence? Are you sure it's not an artifact of the library prep?

5 seconds of googling turns up this:

https://support.illumina.com/content/dam/illumina-support/documents/documentation/chemistry_documentation/experiment-design/illumina-adapter-sequences-1000000002694-13.pdf

This sequence:

ACCACCGAGATCTACA

is on the adapters.

So is this

GACGGCATACGAGAT

Bacterial contamination would not put bacterial sequence at the end of host transcripts! But adapter contamination sure could be.

Here's the index 2 adapter

AATGATACGGCGACCACCGAGATCTACAC[i5]TCGTCGGCAGCGTCAGATGTGTATAAGAGACAG

Here's your sequence

CACCACCGAGATCTACACCTAAGCCTTCGTCGGCAGCGTCAGATGTGTATAAGAGACAGGTATCAACGCAGAGTACGGG

Here's it's blast against nr

ACCACCGAGATCTACACCTAAGCCTTCGTCGGCAGCGTCAGATGTGTATAAGAGACAGGTATC

It's all adapter.

Whoever did the transcriptome assembly didn't trim for adapters first. That's all.

$\endgroup$
6
  • $\begingroup$ I am not sure of anything. I just want to know what are the possibilities. The data was downloaded from the Internet. It's not always the same sequence. For example, I have this two at an extreme of a GAPDH RNA: CACCACCGAGATCTACACCTAAGCCTTCGTCGGCAGCGTCAGATGTGTATAAGAGACAGGTATCAACGCAGAGTACGGG and GACGGCATACGAGATCAGCCTCGGTCTCGTGGGCTCGGAGATTTTGT $\endgroup$
    – juanjo75es
    Jul 12 '20 at 17:16
  • $\begingroup$ Apart from the adapters, how do you explain the presence of the rest of the bacterial sequence? $\endgroup$
    – juanjo75es
    Jul 13 '20 at 19:08
  • $\begingroup$ Sorry swbarnes2 but I do not follow you. If I make BLAST search of that sequence it points to Ralstonia solanacearum with e-value 6e-23, query cover 83% and percentage of identity 100%. I made the BLAST search at blast.ncbi.nlm.nih.gov/Blast.cgi I certainly do not trim anything as I do not need that for assembling the transcriptome but that seems a more than reasonable answer. I just guess now why such long adapter and why using an adapter that matches bacterial dna. Hos do you make that blast search? $\endgroup$
    – juanjo75es
    Jul 13 '20 at 20:58
  • $\begingroup$ Could it be that the assemblies that BLAST finds wrongly include adapters in the assembled sequences? The sequence is repeated a lot of times in the assembly and in other organisms. $\endgroup$
    – juanjo75es
    Jul 13 '20 at 21:37
  • $\begingroup$ You should trim adapters away. Looks like the Ralstonia sequencing group did not. $\endgroup$
    – swbarnes2
    Jul 13 '20 at 22:37

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.