Beyond the sheer volume of sequences that are synthesized and analyzed in a given period, how does the new synthesis pipeline differ from the old one? Are individual DNA templates purchased for each sequence to be synthesized, or has EteRNA moved to a new approach?
In previous labs, SHAPE data for residues near the termini of the molecule are usually not available. Specifically, data for the primer binding site and ~ 10 residues toward the 5’ end of the RNA are usually not shown. For the “player projects,” the primer binding site and 5’ terminus are not shown. If different techniques are being used to synthesize the RNAs and/or probe the structure, is it likely that data will be available for the entire molecule?
Are there plans to continue the FMN aptamer series for several more rounds? Will players be able to design their own sequences that contain the FMN-binding loop? Will other aptamers be investigated?
I’ve heard that there will be in vivo experiments using DMS and CMCT to probe structure of riboswitches. Are there particular switches that will be investigated?
Sorry for the deluge of questions. Any answers are extremely welcome!
Hey, good questions, and I bet some of the other top players have the same ones.
The new ‘cloud biochemistry approach’ involves getting each sequence synthesized, but using various lithography approaches that make 10^3 to 10^6 sequences in parallel; the methods were originally developed to create ‘DNA chips’, a.k.a microarrays.
We will be attaching primer binding sites to player project designs, as well as so-called barcodes necessary for deconvolving the chemical mapping signals from the huge amount of sequencing data. You will see the data for the barcodes, but not for the primer binding site. However we plan to use a protocol where the primers will ‘protect’ the binding site during chemical mapping, so we hope you won’t have to worry about them. And we’re creating special barcodes that shouldn’t interfere with structure; plus you’ll get their chemical mapping data, and we’re planning to have each design coupled to a few different barcodes to help test for systematic errors. We’re actually trying several designs for these ‘flanking sequences’, which is why we didn’t give you a specific aditional 3’ (or 5’) sequence in the player projects – later, once we’ve optimized the protocol we will be quite explicit about these sequences (as well as the entire protocol) to better help you optimize your designs and perhaps even hack the system to do things we haven’t thought of.
Its going to take a few more months to get initial libraries, rigorously optimize, and cross-check the new experimental approaches. We are literally trying all possible sources of DNA now and all possible sequencing methods. Part of the reason we asked for ~1000 sequences now is that we want a good sense of what we’re really going to be making, e.g., are there particular subsequences or terminal sequences that we should avoid.
So for the next few EteRNA play cycles, we will continue FMN stuff – the data are extremely fascinating, as I’ll report in the next couple days – and likely move to switch design soon, but still as global ‘lab puzzles’. When we later transition to the new cloud, we’ll not explicitly test binding of small molecules, but there will be some interesting ways to ‘mimick’ FMN binding, or other aptamers. In fact there are going to be quite a few cool ways to ‘hack the cloud’ to do non-standard stuff, and we’re expecting you to discover many many more. More on that later. Its pretty slick.
My lab is indeed exploring experiments in living cells – by ourselves and also with collaborators – but I can’t give you a timescale for bringing those into EteRNA – they’re definitely in the cards, but right now our focus is on the cloud! [Actually its my focus – I’m the one developing the protocol personally, and its enormous fun.]
Quasi, big thanks for asking these questions. And Rhiju, as you suspected we really liked having the answers. Lots of interesting and exciting things going on.
which is identical to the nupack finger winning solution - so hopefully the various markers and flanking sequences will produce a similar/identical result.
In another post, I had asked Tom if CMCT was going to be integrated into the current EteRNA schedule, and his reply was that at the current time, they are unable to do so. However, my question is, once the new high throughput synthesis pipeline is in place, will we be able to experiment with alternate chemical mapping techniques? I was also reading the synthesis protocols of the Das lab, and saw three other mapping protocols.
ENU (ethylnitrosoururea) mapping for exposed phosphates
OH• (hydroxyl radical) mapping for exposed ribose
iodine/phosphorothioate mapping for dynamic nucleotides
I understand that these still seem to be under development, but if we were to be able to use alternate chemical mapping protocols, would these be available for use as well?
We are actively checking and revisiting these alternative mapping strategies (in fact I ran experiments today on 16 photo-oxidizers). However, its not yet clear whether we will implement them generally for EteRNA solutions.
With the current readout (capillary electrophoresis), each additional mapping method costs about $2 per design – not bad, but it can add up, especially in terms of keeping the experiments organized.
As we transition to a deep sequencing readout, we should be able to do arbitrary numbers of mapping conditions without a big increase in cost or complexity, but at the expense of signal-to-noise (fewer statistics for each mapping condition) or in the number of designs tested in parallel.
The good news is that we’re doing a comprehensive scan of mapping strategies now, and should be able to pick a subset that are most informative (as can be defined quantitatively in bits!). We hope to make a call on which ones to use by early 2013.
I was reading over the original response here, and I had a question about something.
“When we later transition to the new cloud, we’ll not explicitly test binding of small molecules, but there will be some interesting ways to ‘mimick’ FMN binding, or other aptamers. In fact there are going to be quite a few cool ways to ‘hack the cloud’ to do non-standard stuff, and we’re expecting you to discover many many more. More on that later. Its pretty slick.”
So, with the cloud lab nearing launch, I was wondering if it is still going to be the case for RNA switches not needing a small molecule to switch, and if not, what will we be using to cause the RNA to switch instead?
We will test all the sequences without FMN, and then to mimic the fully FMN bound state, we will mutate the FMN binding loops to form Watson-Crick pairs. In fact, by using a series of mutations we should be able to mimic different concentrations of FMN. Well, that’s the hypothesis anyway.
To test this strategy, we will take switches in an upcoming big experimental cycle (probably the second big one, in Feb/March) and test them without FMN, with FMN, and with a couple of these mutations. Should be fun to look at the data!
Oh cool, and I look forward to that data! Will your lab continue to have “lab challenges” where the RNA will switch in the presence of a small molecule once the cloud lab is launched? Or is the next step in that area going to be the in vivo Theophylline switches with Brent Townshend?
But the real question is what do *you* want to pursue? If its a question that can be framed within the new experimental pipeline, you can tackle it. You won’t be reliant on our lab for deciding what questions to ask.
Well, perhaps for testing particular designs in cells you’ll still need help from us or Brent, but who knows, maybe by 2014, that will be part of the EteRNA pipeline too.
Cool, and I look forward to the new pipeline! I do have some ideas I want to test, like just how much stability loop mismatch bonuses provide to synthesized RNA, a very primitive switch strategy, and some anomalies I’ve seen in previous labs, related to both high G-C pair concentration and closing base pairs. I’m sure there are various others I’ve come up with but can’t remember though
