Determine the most likely order of mutations from one RNA sequence to another.

Walter Costa, Maria Beatriz and Hoener zu Siederdissen, Christian and Tulpan, Dan and Stadler, Peter F. and Nowick, Katja  
*Uncovering the Structural Evolution of the Human Accelerated Region 1*  
2017, submitted  
[preprint](http://www.bioinf.uni-leipzig.de/~choener/pdfs/wal-hoe-2017.pdf)  

General information

Given two RNA sequences, one ancestral, and one extant, we want to determine the most likely path of evolution under different measures of fitness.

This program produces the (i) maximum-likelihood path, (ii) all end probabilities, (iii) all start-end probabilities, (iv) all edge probabilities, and (v) the maximum expected accuracy path for these two RNA sequences.

In detail:
(i) gives the optimal path(s) for the fitness function
(ii) gives for each nucleotide polymorphism, how likely it is, that this mutation was introduced last
(iii) looks at all pairs of (first mutation, last mutation) and gives the probability that these two mutations are the begin and end of the chain of mutations
(iv) yields for all pairs of nodes (i -> j) the probability that this path occurs, over the whole ensemble of all possible paths
(v) produces the path of maximal weight using the probabilities produced in (iv)

Usage instructions

sequence generation

First, the sequence data base needs to be created. The following assumptions are being made:

• chimp_118.fa is the origin sequence.
• human_118.fa is the target sequence.
• all known mutations are to be ordered.
• One intermediate (or backmutation) is allowed. This will already lead to an expansion of the sequence space from ca. 250K sequences to 83.6M sequences! Use your local compute cluster or download our precalculated data.

The following command will prepare the working database and populate the seqs subdirectory.

mkdir workdb
mkdir workdb/seqs
mkdir workdb/rnafold
./MutationOrder gensequences -w workdb --ancestral chimp_118.fa -e human_118.fa -g 1 --sequencelimit 100000000 --alphabet=ACGT --seqsperfile=100000

example usage

We assume that you have two Fasta files, chimp_118.fa and human_118.fa but they can be named however is convenient. Each file has to contain exactly one sequence and both sequences have to be of the same length.

For testing with chimp and human, the provided chimp-human.json.gz database should be used, otherwise the initial foldings will be recalculated. All required files are available under 'Binaries' at the bottom of the page.

In case, you don't want or can't use the provided work database, run ./MutationOrder with --verbose

We then run

./MutationOrder --workdb chimp-human.json.gz --scoretype pairdistcen --onlypositive --outputprefix test chimp_118.fa human_118.fa

This will generate test.run, test-edge.eps, and test-meaorder.eps.

The test.run file provides extensive output of the optimal path, the first-last probabilities, the edge probabilities, and the mea output. This conforms to (i) -- (v) mentioned above.

The two eps files give a graphical representation of the edge probabilities, for the meaorder in order of the path of maximum expected accuracy.

The work database collects intermediate structures and their folding and is only created once. The initial run will, however, take some time. I.e. for 'HAR1' this requires 1-4 hours depending on the machine. Further runs complete much faster. In minutes for HAR1.

Command-line options

--help        provides short help
--verbose     will show folding steps during the initial run

-w
 --workdb=ITEM              the database where to store intermediate foldings
-t
--temperature=NUM           annealing temperature. Values close to 0 favor optimal paths. The default is 1.0
--fillweight=FILLWEIGHT     provides logarithmic and linear fill styles for probability plots. The full style always fills the box
--fillstyle=FILLSTYLE       normally, boxes are sized, but all in the same color. This changes the opacity of the color as well. Does not work well for eps files
--cooptcount=INT            how many co-optimals to count for (the count in the .run file is produced differently)
--cooptprint=INT            how many co-optimals to actually print
--outprefix=ITEM            how to prefix all output files
--scoretype=SCORETYPE       choose 'mfe', 'centroid', 'pairdistmfe', or 'pairdistcen' for the evaluation of each mutational step
--positivesquared           square positive energies to penalize bad moves
--onlypositive              minimize only over penalties, not energy gains
--equalstart                each possible mutation is selected with equal probability as the initial one
--posscaled=NUM,NUM         in =x,y will exponentiate all numbers >=x by the constant y. For value k>=x, we have k^y
--lkupfile=ITEM             developer option to feed the initial work database with known foldings (usable but very raw and undocumented. needs 5-line rnafold output)
--showmanual                will show this manual

The allowed score types are:

mfe

which optimizes based on the minimum free energy of each intermediate sequence centroid which instead looks at the energy of the centroid structure pairdistmfe which minimizes the base pair distance between following mutations using mfe structures pairdistcen which minimizes the base pair distance between following mutations using centroid structures

Installation

Pre-built binaries for Linux are avaiable under github releases

Follow this link to the bottom of the page for instructions to build from source.

Contact

Christian Hoener zu Siederdissen
Leipzig University, Leipzig, Germany
choener@bioinf.uni-leipzig.de
http://www.bioinf.uni-leipzig.de/~choener/