Thursday, 18 February 2016

ARTICLE: Who keeps the plans for our power stations?

Evolutionary Inference across Eukaryotes Identifies Specific Pressures Favoring Mitochondrial Gene Retention

IG Johnston, BP Williams
Cell Systems 2 (2), 101-111 (2016)

  • Why some genes are retained in mitochondria, where they are prone to disease-causing mutation, is a much-debated evolutionary question: we use new and generalisable maths and statistics to harness a large volume of sequence data and find the features of genes that predict the patterns of mitochondrial evolution that we observe.

Billions of years ago, a single-celled organism that would become our ancestor engulfed another smaller single-celled organism. The engulfed cell was probably intended to be lunch, but for reasons that remain mysterious (though recently explored here), it remained intact within our ancestor. It produced valuable chemicals that our ancestor could make use of, and was protected within the larger cell. This started a mutually beneficial relationship that evolved over billions of years to give rise to our situation today -- we are the descendants of the big cell, and our mitochondria are the descendants of the small, engulfed cell. 

As they were once independent organisms, mitochondria possess their own genomes (mitochondrial DNA, or mtDNA, which we've written about before). However, unlike the genomes of independent single-celled organisms like bacteria, mtDNA has only a handful of genes: why? Over evolutionary time, the majority of genes have either vanished from mtDNA or been transferred to the nucleus of the host cell. The reasons for transferring these genes to the nucleus are quite well understood; the nucleus is a safer environment for genes, less prone to mutation, and has several other evolutionary advantages.  But, given that transfer to the nucleus is possible, and genes in mtDNA are susceptible to mutation and damage (often giving rise to devastating diseases, which we study and try to prevent), why have mitochondria retained any genes at all? 

This question has been asked for decades, but until recently we lacked the data and the mathematical language to answer it quantitatively. Scientists energetically debate several different hypotheses: our approach attempts to let the data speak for itself without any preconceived ideas about which hypotheses are most likely. To this end, we built a mathematical model encompassing the evolutionary history of organisms with mitochondria, and a powerful statistical framework to amalgamate all the data that has been collected in recent years -- thousands of mitochondrial genomes from organisms from plants to protists (and humans) -- and harness it to compare the many disputed hypotheses addressing this question. 

Our mathematical approach allows us to "rewind the tape of evolution" and explore how mitochondrial genes have evolved. We're looking at Complex I -- an important protein complex involved in respiration -- over time, and watching the number of its subunits encoded in mitochondrial DNA (coloured black) decrease over evolutionary time, according to rules which we identify. The skyscrapers in the background are part of a graph describing how more mtDNA genes are lost over evolutionary history.

In a new paper in Cell Systems here (free here) we found several features that are most related to whether a gene is retained in mtDNA. Before discussing what they were, note that this picture -- several different features each with some influence -- explains and justifies the existing scientific debate. If hypothesis X and hypothesis Y both represent parts of the underlying "truth", then scientists advocating X alone and scientists advocating Y alone are neither completely wrong nor necessarily at odds -- everyone's partly right and the truth lies in the combination of the two arguments. 

The features that predict mtDNA gene retention are how central a gene's product is in its protein complex, the hydrophobicity of the protein the gene encodes, and the proportion of G's and C's in the gene's sequence. This suggests that genes are retained in mtDNA:  
(a) To allow local control of mitochondrial machinery (individual mitochondria can be controlled in response to cellular demands, rather than having to apply changes to the entire cellular population of mitochondria at once).  
(b) To prevent hydrophobic proteins ending up in the wrong place in the cell (if encoded by the far-away nucleus, these proteins may not be able to reach or enter the mitochondrion). 
(c and most speculatively) Because they are capable of withstanding the damaging environment of the mitochondria (GC-rich DNA and RNA is chemically more robust than GC-poor molecules). 

We found that the combination of the features we identified also predicted the success of experiments where scientists have attempted to mimic evolution and artificially transfer genes from the mitochondrion to the nucleus. Our results, as well as addressing a central mystery of evolutionary biology, thus also have the potential to inform synthetic biology approaches to tailor the genetics and bioenergetics of organisms. One final but important point is that the mathematical and statistical machinery we built for this project is highly generalisable and an efficient way of harnessing large sets of data about evolutionary and progressive processes -- we hope to use it to explore lots of other questions, including figuring out the pathways of disease progression and suggesting personalised medicine strategies in the clinic. Iain and Ben

Sunday, 14 February 2016

ARTICLE: Happy Valentine's Day!

Endless love: On the termination of a playground number game

  • A well-known playground game aims to compute a "love score" between two players based on the letters in their names and demonstrates surprisingly rich mathematical behavior; we use some back-of-the-envelope maths and computer simulation to find the highest-scoring names, situations in which the game never ends, and general rules underlying its behaviour.

The "Love Calculator" game is played in playgrounds and online across the world. It computes a playful "love compatibility" by first writing down the counts of l's, o's, v's, e's and s's that appear in two partners' names, then repeatedly adding numbers written next to each other until a percentage score is found -- "Alice loves Bob 54%!" (11010 -> 2111 -> 322 -> 54, as in the picture).

Our light-hearted research project has used maths and computer simulation to explore how the game behaves with different names -- including pairs of all the most common childrens' names in the UK -- and in different languages. "Endless love" -- when the game gets stuck in a loop, or keeps expanding forever -- often occurs between names with high letter counts (like Reese Witherspoon and Calvin Harris).

The project found that any point in the game can be described as a point in a mathematical "space", and that each step in the game moves a point in different ways, like different pieces on a chessboard. While individual outcomes are hard to predict, the average behaviour shows patterns that are repeated over games. The space contains a "cliff"; if moves carry a game over the cliff, it will continue forever.

(left) Alice and Bob playing the "loves" game. The number of l's, o's, v's, e's, and s's in their names give the first string of numbers. Adding neighbouring numbers produces the next strings, until we arrive at the 54% score. But the game never stops for some name combinations. (right) A mathematical picture of the game. w is the length of a string of numbers; m is the sum of the numbers in the string (so 11010 would have w = 5 and m = 3). The arrows show how w and m change on average as a game progresses. The blue region moves left toward an m = 2 final score; the red region keeps growing (or looping), never reaching a final score, and leading to "endless love". 

Different patterns of the "loves" letters give different expected scores. Among the most common childrens' names in the UK, Connor has the highest expected score of 67%, with Evie, Holly, Lola, Molly and Olivia also scoring highly. Names with no "loves" letters -- from Adam to Ryan -- have the lowest expected score of 26%. The most successful names have a middling number of "loves" letters, between 2 and 6. Pairs of o's, several l's, and an absence of some other letters seem to be the key to success -- though a full theory of which patterns give which scores remains elusive.

"Endless love: On the termination of a playground number game" is due to appear in Recreational Mathematics Magazine and is available here. Iain

Wednesday, 3 February 2016

Mitochondrial / math bio conferences

Here's a very rough list of some mathematical biology and mitochondrial conferences that look interesting in 2016!

Generally: Some good listings for math bio conferences; some more

European Conf Mathematical Theoretical Biol 2016Nottingham Jul 11-15 Feb 14
Biomath 2016Bulgaria Jun 19-25 GONE
Int Conf Math BiolPrague Mar 30-31 Jan 31 GONE
Midwest Math bioWisconsin May 21-21Feb 19
Pacific Symposium biocomputingHawaii Jan 4-8 2017??
Algorithms for computational biologyTrujillo Spain Jun 21-23 Feb 2
International symposium bioinf research applicationsMinsk Belarus Jun 5-8 Feb 15
European confernece computational biologyHague Sep 3-7 Mar 29
Systems Biology of Mammalian CellsMunich Apr 6-8 Feb 15
Math mdoelling biology medicineCuba Jun 8-17 Feb 28
Comput Methods Sys BioCambridge Sep 21-23 ??
Applications of Mathamtics to Nonlinear SciencesNepal May 26-29 Feb 28
Information Probability Inference in Systesm BiologyAustria May 18-20 Mar 31
BAMM Biology and Medicine through MathsVirginia May 20-20 Mar 1
Foundations of systems biol in engineeringGermany Oct 9-12 Mar 20
Prac Appl Comp Biol BioinfSeville Jun 1-3 Feb 5

Generally: Some mito confs ;some more

Mitochondrial MedicineCambridge May 4-6 Mar 2
Mitochondrial MedicineSeattle Jun 15-18 ???
CSH Asia: MitochondriaChina Oct 12-16 Aug 21
GRC Mitos and ChlorosVermont Jun 19-24 May 22

Evolution, energetics and noise in the press!

Here are a few times our work on evolution, energetics and noise has appeared in the science and/or popular press, or has otherwise been in the public eye.

The evolution of C4 photosynthesis and design of efficient crops (blog article here)
The Scientist; Science Daily; BANG! Science; eLife; Imperial College; University of Cambridge; UK Plant Sciences Federation;; Innovations Report; e! Science News; Science Newsline 

Evolution of mtDNA gene content (blog article here)
Science (magazine); Science (news); Science's #1 favourite 2016 news article; The Scientist; Cell SystemsEurekAlert; Science 2.0;; Science Daily; MedIndia; University of Birmingham

MtDNA segregation and potential issues with gene therapies (blog article here)
UK HFEA (review here); Science Daily; Imperial College; Health canal; Medical Xpress

Potential issues with gene therapies in real human populations (blog article here)
UK HFEA review 2016

Worldwide vaccine confidence (blog article here)
World Economic Forum, La Vanguardia, Le Monde, International Business Times, Scientific American, Fox News, Vox, Yahoo News, New Scientist, Science, Daily Mirror, Daily Mail 

Caladis, a probabilistic calculator (blog article here)
Nature; Biophysical Journal

"Endless love" and the behaviour of a playground game (blog article here)
University of Birmingham

Mitochondrial "pulsing" in plants (blog article here)
Recommended by F1000
cpYFP responds to pH, not superoxide (blog article here)
Science Daily; Health Medicine Network; EurekAlert; Science 2.0; Health Canal; Medical Xpress; Bioportfolio

Polyomino self-assembly (blog article here)
Our polyominoes made an appearance at the "calcuLate" Science Museum Lates in Nov 2015; simulation website
Videos here; here

Polyominoes modelling proteins (blog article here) 
One of the most-cited articles in Interface for 2014
Virus self-assembly (blog article here)  

Journal front cover and 2010 highlight; video here