Your criteria for voting in the new lab.

We only have 24 slots so it might help narrowing down candidates forum style before your actual
vote. Omei and Jandersonlee put in a lot of time and effort on the project so I would try to make sure their strategies get selected (in my humble opinion). But what about the others?  

1 Like

Excellent question, excellent timing, JR!

I’ll start by say that although a lot of work by a lot of people (not just Jeff and I, and in particular @LFP6 ) has gone into new features for the ribosome challenge, none of us think that we have captured all the relevant aspects, because we don’t know what all the relevant aspects are.

I’ll be dividing my votes between two pretty much distinct categories. One category will go to designs that do well on the “big three” set of criteria - conforming to the soft constraints and finding a good balance between the competing goals of minimizing the number of mutations and minimizing the energy delta.

But I will also be looking for designs that are based on specific hypotheses not reflected in the big three criteria, but for which the designer sets forth an alternate rationale. For example, that might be a specific reason for why it will be informative to test this sequence and what will be learned as a result, whether it be beneficial or harmful to protein transcription.

One criteria I like is to minimizing the energy delta, with and without soft constraints, while minimizing state two structure changes from the original. A sort of: Can we make any effort at lowering the energy delta without blowing the whole ribosome out of the water.


I will probably give most of my votes to designs that meet the soft-constraints chiefly because these are mutations that have been seen in nature in working ribosomes. That said, I will also give some votes to designs that *don’t* stick to the soft-constraints but *do* do a good job of stabilizing stems (less red) and reducing the energy delta. 

I’d also like to see a spread in number of mods, from just a few carefully chosen mods, to large numbers designed to drive down the energy delta. Since we don’t know where the “sweet spot” is here, I think a broad spectrum selection will be good. More delta for less mods *may* be good, but even then some high-delta mods may drop the delta at an expense of unanticipated miss-folds, so voter beware!

I like JR’s criteria on not changing the miss-folding structure elements much (i.e. not adding in new miss-folds) but unless someone creates a Google Doc or forum post comparing alternative designs I probably won’t spend too much time researching that myself.

Lastly I will likely limit myself to one design per player per puzzle so that more players hopefully get a shot at synthesis.

I’ve noticed that in most of my max delta improvement design attempts there seems to be too much structural change from target - which will likely cause function failure - so I will not vote for that type of design.  On 16S and 23S, I will try to vote for designs are closer to key structures…but I don’t really have a good way to do that!  I will filter for soft constraints.

What I do is to take a part of the RNA ( the chicken legs at the bottom of the natural state ) and make that part static in my designs so at least one part of the RNA conforms to the original structure. 

I’m pretty much on the same page as Jeff for voting. I’ll add that any high mutation design should honor soft constraints for reasons already mentioned.

I focused on eliminating global misfolds and stabilizing stems in my designs, which was quite difficult on its own without the added soft constraints. I did back out as many mutations and soft constraint violations as I could for the final 16s and 23s design.

Here is my voting strategy:

  1. First I vote for designs with very few mutations. Here is why. The E. coli ribosome already works as is, but each additional base mutation exponentially increases the risk of destroying something. Plus few mutations will also make it easier when results comes back, to detect which of the changes were positive or not. This will also make it easier to pool positive mutations in a later round.   

  2. I primarily look for designs that follow the soft constraints, although I may dispensate if the change is made in a region of my interest.

  3. I look for designs that make changes in regions that I think will have a strong effect on the design. These are:

  • Closing basepairs in multiloops (#multiloopGCorientation)

I’m interested in single base pair change in multiloops closings - conserved or not. Specific orientation of closing base pairs in multiloops affect stability of static RNA designs. When a puzzle grows really big, not all closing basepairs can be GC or be of the otherwise most optimal orientation, as this will cause too much repetitive sequence and thus result in misfolds and switching instead.

  • Breaking up strong repetitive sequences (#sequencevariation)

Too much base repeat make static design unstable and switchy. While some localized regions in the ribosome is switching, not all regions need to be.

  • Neck (#neck)

I believe it is possible to alter the overall stability of an RNA by altering its neck sequence. This only holds relevance to the 5S and 23S rRNA. I will be similarly interested in mutations in “subnecks” neck regions for the individual partials in 16S and 23S.  

  • #Non-A base in single base area

I’m interested in single non-a base change in single base area - for bases that are not conserved. In past static RNA labs, too much and strong non A loop sequence in loop areas, increased the risk of misfolding. Especially if there was matching sequence elsewhere in stems also.

  • GU and GU mismatches (#GUmismatch +/-)

I’m interested in GU deletion and GU addition plus GU mismatches in particular. I think they have a role for switching in the ribosome. I prefer change where conservation suggest it to be successful or if from a more distant species as well.

Usually I go for the data. Either the Dot plot and the melt plot. Like the guide says, if in doubt go for the dot plot. I will admit though usually I go for the constraints. Usually the most clear without any red what so ever. as for the energy levels, I have a hard time having good knowledge on what they represent and exactly how it the levels work, so I don’t read into it too much. 
over all though I go with the least amount of red. ( arc plot engines in Eterna would be nice to look at while looking at other players labs)

There is an Arcplot booster in the scripts if you haven’t already made yourself a copy.

Top 13 contenders for 5S to help assess design variation:
JL     A12,A26,C65,A87
Eli     A12,A20,U41,U63
Korosi U41,C108
rhiju   U35,A76,U88
Omei  A26,A76,A87
Jieux   A12,A36
Gerry  A18,U107
MS     A12,A25,A87
Astr3  A12,C41,A107
Astr4  A12,A20,U63,A87
nik      A12,A13,U41,C87,U88,C108
GS4   A25,A76,A87
Jieux  A11,A35