How To Detect Horizontal Gene Transfer?

Question

[Aleksandr Levchuk]

Horizontal gene transfer in Bacteria can occur by transformation, transduction, bacterial conjugation.

Experiment outline:

• Two bacterial communities are mixed together and left thriving for some time (e.g. 24 hours)
• An SMRT (single molecule) sequencing is performed on only 1 sample (the mix of the 2 communities)

Assumptions (1st set):

1. The transfers can be genes or a some other part of the genome
2. The transfered regions are large enough and unique enough to be clearly distinguishable
3. The combined community contains a large amount of donors, recipients, and siblings of the recipients that are not recipients

How would you approach the bioinformatics for finding the horizontal gene transfers?

How can one distinguish between ancient transfers and recent transfers?

(Any literature links or tools would be appreciated.)

January 17th addition:

To illustrate what I'm imaging, I offer the following multiple alignment which I put together.

                                   20                  40
                 ~~~~~~~~~~~~~~~~~~~+~~~~~~~~~~~~~~~~~~~+~~~~~~~~~~~~~~~~~
dna-molecule-01  TAGGGGCACGATCATGCCTTAGTTTTTTTGCGACTGTCGCTCTTTCGCGTCTCAGTC
dna-molecule-02  TAGGGGCACGAACATGCCTTATTTTTTTTGCTACTTTCGCTCTTTTGCGTATCAGTC
dna-molecule-03  TAGGGGCACGATCATGCCTTAGTTTTTTTGCGACTATCGCTCTTTCGCGTCTCAGAC
dna-molecule-04  TAGGGGCACGATCATGCCTTAGTTTTTTTGCGACTGTCGCTCTTTCGCGTCTCAGTC
dna-molecule-05  TAGGGGCACGATCATGCCTTAGTTTTTTTGCGACTGTCGCTCTTTCGCGTCTCAGTC
dna-molecule-06  TAGGGGCACGAACATGCCTTATTTTTTTTGCTACTTTCGCTCTTTTGCGTATCAGTC
dna-molecule-07  TAGGGGCACGATCATGCCTTAGTTTTTTTGCGACTATCGCTCTTTCGCGTCTCAGAC
dna-molecule-08  TAGGGGCACGAACATGCCTTATTTTTTTTGCTACTTTCGCTCTTTTGCGTATCAGTC
dna-molecule-09  TAGGGGCACGATCATGCCTTAGTTTTTTTGCGACTATCGCTCTTTCGCGTCTCAGAC
dna-molecule-10  TAGGGGCACGATCATTCCT---------------------------GCGTATCGGAC
dna-molecule-11  TAGGGGCACGATCATGCCT---------------------------GCGTATCAGAC
dna-molecule-12  TAGGGGCACGATCATGCCT---------------------------GCGTATCAGAC
dna-molecule-13  TAGGGGCACGATCATTCCT---------------------------GCGTATCGGAC
dna-molecule-14  TAGGGGCACGATCATTCCT---------------------------GCGTATCGGAC
dna-molecule-15  TAGGGGCACGATCATGCCT---------------------------GCGTATCAGAC
dna-molecule-16  TAGGGGCACGATCATGCCT---------------------------GCGTATCAGAC
dna-molecule-17  TAGGGGCACGATCATGCCT---------------------------GCGTATCAGAC
dna-molecule-18  TAGGGGCACGATCATGCCT---------------------------GCGTATCAGAC

More assumptions (2nd set):

1. A pattern above is actually found in the sequencing data
2. The sequencing technology guarantees that each sequence comes from 1 individual only
3. All 18 molecules are shown to belong to the same progeny A
4. The "TAGTTTTTTTGCGACTATCGCTCTTTC" sequence or anything similar is not observed anywhere in the the genome of the progeny A, but found in most of the genomes of a different progeny B from the same sample.

If the above pattern is observed in vio would this be strong evidence of horizontal gene transfer?

Which of the above assumptions are not reasonable?

[Lars Juhl Jensen]

I am not sure I understand your experiment and the goal. The two cultures contain bacterium A and B, respectively? You then mix them, wait some time and sequence? And the hope is to count the frequency with which transfer of DNA takes place between A and B?

[Michael]

I am concerned that experiment wouldn't work.

[Aleksandr Levchuk]

@Lars, that's pretty much what I had in mind but instead of 1 bacterium A and another bacterium B, would prefer 1 bacterial community A and another bacterial community B. My final goal is to do Microbial Forensics. For this experiment I would only want to be able to see some specific cases where the gene transfers took place, what sequences where transfered, and some gnomic context of the transfered regions.

[Aleksandr Levchuk]

@Lars, that's pretty much what I had in mind but instead of 1 bacterium A and another bacterium B, would prefer 1 bacterial community A and another bacterial community B. My final goal is to do Microbial Forensics. For this experiment I would only want to be able to see some specific cases where the gene transfers took place, what sequences where transfered, and some genomeic context of the transfered regions.

[Aleksandr Levchuk]

Bacterial Transformation (a Flash Animation)

Answer 1

[biobot 0.0.77.a.1099]

The Alien hunter package has been used to detect horizontal transfer in Salmonella, given the complete genome sequences of two species.

However, I have some concerns about your experiment. What is your control? If you detected a putative recent transfer between two species, how do you distinguish it from another pre-existing transfer? The genomes of bacteria are a patchwork of recent and ancient transfers, including prophage and other mobile genetic elements.

It seems the constituents of your two cultures are not known and you are preparing a single, mixed library for short read sequencing. Assuming that a read does align to one (or several) known genomes, you have lost the original genomic context of that 100bp sequence. By that I mean you don't know whether it was transferred or not, or what was next to it in the donor or recipient.