Results: Haplotype Inference by Pure Parsimony Tools

Accuracy Database

Efficiency Database

 

Summary of the performance of HIPP tools

  Last Update: March 2010

This table presents the summary of the performance of each HIPP solver (RTIP, HAPAR, PolyIP, HybridIP, SHIPs, RPoly, Haplo-ASP, HaploPPH, Satlotyper) on a set of 1183 problem instances. "# Solved Instances" states the number of instances solved within 1000 seconds. "# Aborted Instances" states the number of instances which exceeded the time or the memory available. "# Wrong Results" states the number of instances proved to provide a incorrect number of haplotypes. "Average Time" states the mean time required by the solver, on the set of solved instances within the time limit. Results were obtained on an Intel Xeon 5160 server (3.0GHz, 4GB RAM) running Red Hat Enterprise Linux WS 4. Time is limited to 1000 seconds and memory is limited to 3GB RAM.
# Solved Instances # Aborted Instances # Wrong Results Average Time (sec)
RPoly v1.2 1165 18 0 3.53
RPoly v1.1 1147 36 0 6.38
SHIPs 1116 67 0 7.86
HaploPPH 936 247 0 42.34
Haplo-ASP 869 310 4 30.18
RTIP 805 378 0 6.78
Satlotyper 791 392 0 24.41
HAPAR 579 603 1 39.90
PolyIP 476 707 0 73.44
HybridIP 469 714 0 72.94

 

 

 

Implementation

This table presents the language implementation and the solver used by each HIPP method.

HIPP model Implementation Solver
RPoly C MiniSat+
SHIPs perl MiniSat
HaploPPH mosel Xpress-Mosel
Haplo-ASP perl CModels, MiniSat
RTIP C++ CPLEX
Satlotyper java MiniSat
HAPAR C++ (none)
PolyIP C++ CPLEX
HybridIP C++ CPLEX

 

  All tools were provided by the authors,except the RTIP tool. This tool was provided by the authors of PolyIP and HybridIP. To the best of our knowledge, the author of RTIP has not made the software available. HAPAR does not use any constraint solver, because HAPAR is a dedicated branch-and-bound method to solve the HIPP problem.