3 ms·

Short URL to the patent at USPTO: http://1.usa.gov/plfTg3 http://1.usa.gov/plfTg3 I am not a specialist in protein folding, so take it with a grain of salt, bu

Short URL to the patent at USPTO: http://1.usa.gov/plfTg3 http://1.usa.gov/plfTg3 I am not a specialist in protein folding, so take it with a grain of salt, but from the figures it does not look like method predicts actual configurations decoded from experiments. Patent itself acknowledges that (5.50): "the lack of computational prediction [...] can be attributed to problems arising form calculating molecular mechanics potentials (force-fields)". Could anybody knowledgeable make an assessment how important this invention is?

hugh3 15y ago
My knowledge of protein folding algorithms is pretty limited, and in any case everything in this press release is pretty vague. However, this press release is dated September 14, and this is the first I'm hearing about it. A google news search for "Oak Ridge protein folding" shows only a teensy bit of small-time coverage (i.e. reprinting of the press release). I'm inclined to think that if there was a big deal here, somebody knowledgeable would be making more noise about it.
jonmc12 15y ago
I'm no expert, but I have explored his problem in the past as I understand that protein folding simulations are considered by some to be the constraint of biological life extension. Protein folding is an intractable N-body simulation problem that relies on number of different methods to approximate the 3D state of the protein throughout its folding sequence. In reviewing the patent, it looks like this is an incremental improvement to simulation based off some findings in recent years, but does not seem to be any kind of breakthrough in understanding of proteins. Something called 'Vibration Mode' seems to be an empirical observation that explains how the internal dynamic of the protein can limit the number of future 3D states based on a previously observed folding cycle, and a model of the local energy minima and maxima.
jlmelville 15y ago
This paper is describing a way of analysing the output of molecular dynamics simulations of proteins, a series of 'snapshots' of the positions of the atoms in the protein over time. Given the number of atoms in the molecule and the short time steps used, it's useful to be able to identify which parts of the protein (and when) in the simulation are doing something interesting and separate those from the random to-ing and fro-ing that you'd expect to occur as thousands of atoms move about. That's what this paper is describing. It's not being applied to experimental data, and is one of many techniques used to analyse MD trajectories. I'm not an expert in these methods to assess how novel or useful this paper is compared to the others, but I would not expect that it's of interest to a layman.

hugh3 15y ago

My knowledge of protein folding algorithms is pretty limited, and in any case everything in this press release is pretty vague. However, this press release is dated September 14, and this is the first I'm hearing about it. A google news search for "Oak Ridge protein folding" shows only a teensy bit of small-time coverage (i.e. reprinting of the press release). I'm inclined to think that if there was a big deal here, somebody knowledgeable would be making more noise about it.

jonmc12 15y ago

I'm no expert, but I have explored his problem in the past as I understand that protein folding simulations are considered by some to be the constraint of biological life extension. Protein folding is an intractable N-body simulation problem that relies on number of different methods to approximate the 3D state of the protein throughout its folding sequence. In reviewing the patent, it looks like this is an incremental improvement to simulation based off some findings in recent years, but does not seem to be any kind of breakthrough in understanding of proteins. Something called 'Vibration Mode' seems to be an empirical observation that explains how the internal dynamic of the protein can limit the number of future 3D states based on a previously observed folding cycle, and a model of the local energy minima and maxima.

jlmelville 15y ago

This paper is describing a way of analysing the output of molecular dynamics simulations of proteins, a series of 'snapshots' of the positions of the atoms in the protein over time. Given the number of atoms in the molecule and the short time steps used, it's useful to be able to identify which parts of the protein (and when) in the simulation are doing something interesting and separate those from the random to-ing and fro-ing that you'd expect to occur as thousands of atoms move about. That's what this paper is describing. It's not being applied to experimental data, and is one of many techniques used to analyse MD trajectories. I'm not an expert in these methods to assess how novel or useful this paper is compared to the others, but I would not expect that it's of interest to a layman.