This protein from puzzle 1297 was the first player design to be grown, crystallized, and solved by the Foldit science team. Anthropic Dreams players Waya, Galaxie, and Susume contributed to this design. Waya provided this image. The protein was dubbed Foldit1.
One of the major goals of Foldit design puzzles is creating artificial proteins that fold up on their own the way natural ones do.
The Foldit science team has been testing player designs in various ways, and refining the game to produce more realistic proteins.
Players have produced at least four proteins that have passed all the science tests. The proteins have been grown in bacteria, extracted and crystallized, then "solved" using X-ray diffraction. The same process is used to analyze natural proteins.
See Solved Foldit designs for a quick summary of the successful designs, as described in the Foldit protein design paper in Nature.
The first successful player design was produced for Puzzle 1297 by Anthropic Dreams, with Waya, Galaxie, and Susume contributing. The players involved have dubbed this protein Foldit1, and this name is used in the Nature article.
The successful protein was then presented in Puzzle 1381 as a de-novo. The science team finally shared the good news, and presented the protein yet again in Puzzle 1384 with the electron density included.
Results from Puzzle 1452 provided by Susume. The protein, known as Ferredog-Diesel, was designed by tokens in Puzzle 1313.
The second successful design was created by tokens in Puzzle 1313, and named Ferredog-Diesel. It was presented as a de-novo in Puzzle 1452, and as an electron density in Puzzle 1475.
Foldit3, as shared by Susume in the Puzzle 1322 results.
The third successful design, dubbed Foldit3, was created by Susume in puzzle 1322. Foldit3 appeared again as a de-novo in Puzzle 1531. Unlike the other solved Foldit designs, Foldit3 has not appeared as an electron density puzzle.
Beta Folders' evo solution to Puzzle 1511, showing the Peak6 protein.
The fourth successful protein was created by fiendish_ghoul in puzzle 1331. The protein is named Peak6. The protein was presented as a de-novo in puzzle 1511, and as an electron density in puzzle 1516.
There may be more player designs in the wet lab pipeline.
Testing process[]
Graph of Rosetta@home results for a protein designed by a Foldit player. The red dots have a "funnel" shape, converging toward the lower left-hand corner. This shape indicates the protein is likely to fold up on its own. An image of the protein has been superimposed on the plot.
The Foldit blog post The Baker Lab tests Foldit player-designed proteins describes the basic process, which involves testing promising solutions using Rosetta@home and other tools as a first step.
Designs which pass initial tests move on to "wet lab" testing, which involves growing the protein in E. coli bacteria. The blog post The story of a Foldit design details the steps involved in growing, extracting, and purifying a protein.
Protein designs which make it through all the stages of wet lab testing go on to crystallization trials. Proteins don't form crystals easily, so this phase usually involves automated testing of many different methods and conditions. As described in the Foldit blog post Foldit design update - Part 2, "protein crystallization is a lottery".
The Foldit blog post protein crystallization showed an example of a protein that appeared on its way to growing a good crystal.
Once a suitable crystal is grown, the next step is X-ray crystallography to produce an electron density of the protein. The electron density information allows the three-dimensional structure of the protein to be determined, meaning the protein has been "solved".
The Foldit blog post X-ray Diffraction covers this step, and describes the successful solution of the Anthropic Dreams protein from Puzzle 1297.
As was done from the successful protein from Puzzle 1297, proteins designed by Foldit players may be presented de-novo puzzles, in part to supplement Rosetta@home testing.
Successful designs may also appear as electron density puzzles, again as was the case for the Puzzle 1297 protein.
Foldit blog posts on design puzzle results[]
These blog posts from bkoep (Brian Koepnick) and beta_helix (Firas Khatib) describe the results of various Foldit design and electron density puzzles.
- X-ray Diffraction - successful solution of a protein designed by Foldit players (bkoep, 30 May 2017)
- Protein crystallization - the crystal-growing phase of protein analysis (bkoep, 15 April 2017)
- Foldit design update - Part 2 - more details on the testing of proteins designed by Foldit players (bkoep, 1 March 2017)
- Foldit design update - Part 1 - 14 Foldit player designs were successfully folded by Rosetta@home (bkoep, 28 February 2017)
- Player designs enter the wet lab (bkoep, 10 February 2015)
- The story of a Foldit design (bkoep, 16 June 2014)
- Improvements in Foldit designs, with examples of "funnels" seen in evaluating design results (bkoep, 25 March 2014)
- "The Baker Lab tests Foldit player-designed proteins" (bkoep, 1 November 2013)
- More Exciting Electron Density Results! (beta_helix, 8 August 2013)
Foldit trophy![]
In 2009, Puzzle 187 challenged players to redesign human fibronectin (RCSB 1FNF). Based on initial testing, one player design was selected for the "web lab". Things didn't work out quite as hoped, but the player, Boots McGraw, got a unique souvenir.
The following blog posts detail the process:
- Puzzle 187, launched 14 August 2009, inviting players to redesign human fibronection to produce a more tightly-packed core.
- Will this work? We'll see.... describing the testing of one solution to human fibronectin puzzle. (Possu, 16 October 2009
- The first round of results on designed fibronectin are in.... but sadly, the selected fibronectin design didn't do so well in the wet lab. (possu, 22 December 2009)
- Foldit trophy! Boots McGraw was awarded with a trophy consisting of a 3D-printed version of his solution to puzzle 187. (Seth Cooper, 5 April 2010)