Design track observations

This is meant as handwaving explanation of why I think top7 (1qys-A) sounds different from 1ct7 or any of the thousands of proteins I've heard. At the same time, there is some additional free music here where I am recording songs using both the synthesiser and drum modukes and comparing and contrasting. I hope you enjoy!


I have a bunch of observations about how this novel designed protein is different sounding from naturally occuring proteins, and also provide some links to sound files that compare and contrast the two sound modules I'm using.

You can aurally hear some features of top7 that are not present in normal proteins (this observation is based on listening to tens of PDB files thousands of times with different parametres so like 3D structure visualisation it's a skill acquired by listening to structures in many ways). First, it is relatively far more repetitive than natural proteins in terms of patterns (you can hear the same sequence motifs being repeated throughout the songs). Second, the sequence is relatively more simple than natural proteins (this is partly because of the repetition but even when there're no repeated passages it has a rather simplistic overall flow with few variations. Third, there's a symmetric feel to it (shades of Bach) and I bet there is some symmetry in the structure also (this is seen in some natural proteins; I wonder if there's a connection to stability which top7 is optimised for). Fourth, there are some surprises and if you're patient a bit you'll hear it coming; particularly the "guitar solo" portion (which reflects RAPDF scores between all atoms) is more complex than what I've heard in natural proteins (which would make sense given that top7 is really a creation of really good side chain packing). (I'm making all these observations without visually looking at the structure but I have seen top7 before extensively when it was first published.) This track is recorded with the "Sci Fi" drum kit and GM sounds in the Yamaha DTX 900 drum module.

So some extra musical recordings for context and contrast: the translation (1ctf) track from the DTX 900 with a standard kit. You can also contrast this to the original translation and twisted helices tracks recorded using the Yamaha Clavinova CVP 305 synthesiser again with a standard kit and the GM sounds (this is normalised, equalised, and compressed---something I didn't do for the above two tracks so the "apparant loudness" is increased making it sound better than it really is but also muddier).

The simplicity in top7 is a top7 thing, I believe, in a quest for stability. Listen to the fake designer protein twisted helices (which is just the logo PDB file converted into a 20 second song). The complexity of just 20 amino acids is pretty high (the "complexity" in this case is probably more "noise" or unchecked heterogeneity). That same twisted helices track with the DTX 900 sounds (with a club kit) is interesting.


TWISTED HELICES || Ram Samudrala || [email protected]