maxentpy

maxentpy is a python wrapper for MaxEntScan to calculate splice site strength.

It contains two functions. score5 is adapt from MaxEntScan::score5ss to score 5’ splice sites. score3 is adapt from MaxEntScan::score3ss to score 3’ splice sites. They only use Maximum Entropy Model to score.

Prerequisites

• Cython
• msgpack-python

Examples

>>> from maxentpy import maxent  # use normal version of maxent
>>> maxent.score5('cagGTAAGT')  # 3 bases in exon and 6 bases in intron
10.858313101356437
>>> maxent.score3('ttccaaacgaacttttgtAGgga')  # 20 bases in the intron and 3 base in the exon
2.8867730651152104
>>> from maxentpy.maxent import load_matrix5, load_matrix3  # preloading matrix will speed up
>>> timeit maxent.score5('cagGTAAGT')
10 loops, best of 3: 23.5 ms per loop
>>> matrix5 = load_matrix5()
>>> timeit maxent.score5('cagGTAAGT', matrix=matrix5)
100000 loops, best of 3: 3.27 µs per loop
>>> timeit maxent.score3('ttccaaacgaacttttgtAGgga')
1 loop, best of 3: 259 ms per loop
>>> matrix3 = load_matrix3()
>>> timeit maxent.score3('ttccaaacgaacttttgtAGgga', matrix=matrix3)
10000 loops, best of 3: 103 µs per loop
>>> from maxentpy import maxent_fast  # fast version could further speed up
>>> timeit maxent_fast.score5('cagGTAAGT')
100 loops, best of 3: 5.04 ms per loop
>>> timeit maxent_fast.score3('ttccaaacgaacttttgtAGgga')
100 loops, best of 3: 9.3 ms per loop
>>> from maxentpy.maxent_fast import load_matrix  # support preloading matrix
>>> matrix5 = load_matrix(5)
>>> timeit maxent_fast.score5('cagGTAAGT', matrix=matrix5)
100000 loops, best of 3: 3.61 µs per loop
>>> matrix3 = load_matrix(3)
>>> timeit maxent_fast.score3('ttccaaacgaacttttgtAGgga', matrix=matrix3)
100000 loops, best of 3: 7.76 µs per loop

Benchmark

score5

score5	maxentpy.maxent	maxentpy.maxent_fast
without matrix	23.5 ms	5.04 ms
with matrix	3.27 µs	3.61 µs

score3

score3	maxentpy.maxent	maxentpy.maxent_fast
without matrix	259 ms	9.3 ms
with matrix	103 µs	7.76 µs

Citation

Yeo G, Burge CB. Maximum entropy modeling of short sequence motifs with applications to RNA splicing signals. Journal of Computational Biology. 2004, 11:377-94.

License

The original algorithm and perl scripts are under license described in http://genes.mit.edu/burgelab/maxent/download/READTHIS. The python version of maxent is under the MIT License.