Sequence submission, ho!

link

Hi blog! Long time, no update…

On Thursday, I submitted 91 genomic DNA samples to our local genome centre (center for you American reader). They’ll be sheared up, converted into indexed (barcoded) libraries, and sequenced on an Illumina GA2 over 4 “lanes”. Whew! That should have been easier.

I won’t get the data until probably around February, so with that out-of-the-way for now, I’d best get my act together. But before that, here’s what this first round of sequencing for our Genome BC grant is about:

We want to get a basic genetic picture of what happens when we do our standard transformation experiments; we lack simple rules of genetic transmission by natural transformation. How many DNA fragments do competent cells add to their chromosomes by homologous recombination? How long are the segments? How does genetic divergence affect transformation?

link

Haemophilus influenzae KW20 cells can be made naturally competent by starvation of log-phase cells in M-IV media. A fraction of the cells in the culture becomes competent, able to take up added DNA fragments into their cytosol (in H. influenzae, uptake is biased to fragments containing USS, abundant motifs in H. influenzae chromosomes). Taken up DNA fragments can recombine with the cells’ chromosomes, if sufficient sequence identity between the fragment and part of the chromsome allows it.

We added genomic DNA from an antibiotic-resistant derivative of a divergent clinical isolate NP (86-028NP NovR NalR) to competent KW20 cultures in three separate experiments. In each case, we isolated antibiotic-resistant and –sensitive clones, grew these up in overnight cultures, saved frozen stocks, and extracted genomic DNA from the rest of the cultures for sequencing.

In these experiment, we are not getting a particularly “natural” view of natural transformation, since the donor DNA we’re using is purified chromosomal DNA from a single divergent isolate. Since we know effectively nothing about the actual DNA in the natural environment of naturally competent H. influenzae—its source, size-distribution, or associated muck—this seems like a good place to start. We know the genome sequence of the donor and can use the genetic differences between donor and recipient to identify transforming DNA segments.

Based on our preliminary work, we think that these transformed clones will contain long(ish) segments from NP, replacing segments of the KW20 chromosome. We’re using multiplexed Illumina GA2 paired-end sequencing to identify recombinant segments in a large set of clones from these three experiments. We won’t get perfect genome sequences out the other end; instead we expect a median read depth of ~60 per clone, enough to probably capture the vast majority of diagnostic polymorphisms in each clone. Then we’ll look at the distribution of recombinant segments across the independent transformants. That’ll be the Science part.

Here’s what the 91 samples were:

Selected set: 72
Because we estimate that only about 10% of the cells in each culture was actually competent, we selected for either novobiocin or nalidixic acid resistant clones (3x12 of each) to ensure that the clone was derived from a competent cell. To allow for selection for transformed cells, the NP donor DNA was purified from a clone made NovR and NalR by PCR-mediated transformation. This means that for a given selected clone, a recombinant segment is always predicted that spans the appropriate locus. At the two selected loci, we will be able to ask about “LD decay” away from the selected positions. We also expect a large number Independent recombinant segments in these clones; these will provide basic information about the distribution, size, and breakpoints of recombination tracts in natural transformants.

Unselected set: 8 pools of 2
While we estimate percent-competence by evaluating the “congression” (unexpectedly high co-transformation) of “unlinked” markers (not on the same DNA fragment), this calculation relies on an assumption about competent cultures, namely that cells come in only two flavors: either non-competent, unable to take up DNA, or competent, able to take up several long fragments of DNA. It could also be that “congression” arises from a more quantitative distribution of states, where cells are more or less competent; i.e. able to take up and recombine a variable number of fragments.
For the most part, the use of “congression” to evaluate the overall competence of a culture is probably valid, but in terms of generalizing from the above large set of selected transformants, we’d like to know how well this assumption holds true. Because we expect that most unselected clones will look just like the recipient KW20 chromosome, we pooled 16 unselected clones into pairs and include these pools as 8 of our submitted samples. Since we predict that only 10% of cells in the cultures was competent, our null hypothesis is that 1-3 clones will look like selected clones (i.e. a handful of independent recombination tracts) and the rest will look just like KW20. We may, however, find that many more clones show evidence of transformation, but containing only one or only very short recombination tracts.

Selected late-log set: 8
Similarly, we also included 8 NovR clones selected from a late-log transformation. Cultures of cells in late-log are considerably less competent than MIV cultures (~1-2 orders of magnitude), and Rosie showed me some old data in her notebook suggesting that this could roughly be accounted for by percent-competence. The suggestion is that the low transformation frequency of late-log cultures vs MIV cultures is only because fewer cells are competent, rather than late-log competent cells being less transformable. Sequencing a handful of transformants from a late-log culture will help sort this out, in a similar fashion as sequencing unselected clones from MIV cultures.

Controls: 3
We included both the donor and recipient genomes as samples; while we’ve already obtained ridiculous coverage of these samples, including them here acts as an internal “coverage control”. We also threw in some MAP7 DNA, which is a KW20-derivative with a bunch of antibiotic-resistances. While it’s going to look almost like KW20, we use it on a practically daily basis, so it’d be nice to see its sequence.

So that’s the set of 91 clones we’re having sequenced with our first round of Genome BC cash. I’m keeping my fingers crossed that the data is good and abundant and reveals new things…