shortestPath.cc

/*

  Copyright (C) 2003-2009 Kevin Thornton, krthornt[]@[]uci.edu

  Remove the brackets to email me.

  This file is part of libsequence.

  libsequence is free software: you can redistribute it and/or modify
  it under the terms of the GNU General Public License as published by
  the Free Software Foundation, either version 3 of the License, or
  (at your option) any later version.

  libsequence is distributed in the hope that it will be useful,
  but WITHOUT ANY WARRANTY; without even the implied warranty of
  MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  GNU General Public License for more details.

  You should have received a copy of the GNU General Public License
  long with libsequence.  If not, see <http://www.gnu.org/licenses/>.

*/

#include <Sequence/shortestPath.hpp>
#include <Sequence/PathwayHelper.hpp>
#include <Sequence/Translate.hpp>
#include <Sequence/Grantham.hpp>
#include <Sequence/SeqConstants.hpp>
#include <Sequence/SeqAlphabets.hpp>
#include <Sequence/Comparisons.hpp>
#include <algorithm>

namespace Sequence
{
#ifndef DOXYGEN_SKIP //doxygen should skip this
  class shortestPath::shortestPathImpl
  {
  public:
    shortestPath::pathType _type;
    Sequence::Grantham distances;
    double _distance;
    std::string t1,t2;
    std::vector<std::string> _path;
    std::pair<double,shortestPath::pathType>process_path(const std::string &intermediate,
                                                         const Sequence::GeneticCodes & code);
    std::pair<double,shortestPath::pathType>process_path2(const std::string &intermediate1,
                                                          const std::string &intermediate2,
                                                          const Sequence::GeneticCodes & code);
    shortestPathImpl(const std::string &codon1,
                     const std::string &codon2,
                     const Sequence::GeneticCodes & code);
  };

  shortestPath::shortestPathImpl::shortestPathImpl(const std::string &codon1,
                                                   const std::string &codon2,
                                                   const Sequence::GeneticCodes & code):
    _type(shortestPath::pathType::AMBIG),
    distances(Sequence::Grantham()),
    _distance(0.)
  {
    if(codon1.length()!=3 || codon2.length() != 3)
      {
        throw std::runtime_error("Codons are not both of length 3");
      }
    t1 = Sequence::Translate(codon1.begin(),codon1.end(),code);
    t2 = Sequence::Translate(codon2.begin(),codon2.end(),code);

    _path.push_back(codon1);

    //We have already thrown an exception if codons are not the same length
    decltype(Sequence::NumDiffs(codon1,codon2)) ndiffs = Sequence::NumDiffs(codon1,codon2);

    if (ndiffs == 0)
      {
        _type = shortestPath::pathType::NONE;
      }
    else if (t1 != "X" && t2 != "X" && 
             (std::find_if(codon1.begin(),codon1.end(),
                           ambiguousNucleotide()) == codon1.end()) &&
             (std::find_if(codon2.begin(),codon2.end(),
                           ambiguousNucleotide()) == codon2.end()) )
      {
        if (ndiffs == 1)
          {
            _distance = distances(t1[0],t2[0]);
            if(t1 != t2)
              {
                _type = shortestPath::pathType::N;
              }
            else
              {
                _type = shortestPath::pathType::S;
              }

          }
        else if (ndiffs == 2)
          {
            auto intermediates = Sequence::Intermediates2(codon1,codon2);

            std::vector< std::pair<double, shortestPath::pathType> > 
              paths(2,std::pair<double, shortestPath::pathType>(0.,shortestPath::pathType::AMBIG));

            //update pathway lengths
            paths[0] = process_path(intermediates[0],code);
            paths[1] = process_path(intermediates[1],code);
            if (paths[0].first < paths[1].first)
              {
                _type = paths[0].second;
                _distance = paths[0].first;
                _path.push_back(intermediates[0]);
              }
            else
              {
                _type = paths[1].second;
                _distance = paths[1].first;
                _path.push_back(intermediates[1]);
              }
          }
        else if (ndiffs == 3)
          {
            auto intermediates = Sequence::Intermediates3(codon1,codon2);
            std::vector< std::pair<double, shortestPath::pathType> > 
              paths(6,std::pair<double, shortestPath::pathType>(0.,shortestPath::pathType::AMBIG));
            //there are 6 paths b/w codons that differ at all 3 sites
            //the indexing of array intermediates is describing in the documentation 
            //for class Sequence::ThreeSubs in the manual
            paths[0] = process_path2(intermediates[0],intermediates[1],code);
            paths[1] = process_path2(intermediates[0],intermediates[2],code);
            paths[2] = process_path2(intermediates[3],intermediates[4],code);
            paths[3] = process_path2(intermediates[3],intermediates[5],code);
            paths[4] = process_path2(intermediates[6],intermediates[7],code);
            paths[5] = process_path2(intermediates[6],intermediates[8],code);
              
            size_t shortest = 0;
            double shortest_path = paths[0].first;
            for(unsigned i = 0 ; i < 6 ; ++i)
              {
                if (paths[i].first < shortest_path)
                  {
                    shortest=i;
                    shortest_path=paths[i].first;
                  }
              }
            _type = paths[shortest].second;
            _distance = paths[shortest].first;
            switch (shortest)
              {
              case 0:
                _path.push_back(intermediates[0]);
                _path.push_back(intermediates[1]);
                break;
              case 1:
                _path.push_back(intermediates[0]);
                _path.push_back(intermediates[2]);
                break;
              case 2:
                _path.push_back(intermediates[3]);
                _path.push_back(intermediates[4]);
                break;
              case 3:
                _path.push_back(intermediates[3]);
                _path.push_back(intermediates[5]);
                break;
              case 4:
                _path.push_back(intermediates[6]);
                _path.push_back(intermediates[7]);
                break;
              case 5:
                _path.push_back(intermediates[6]);
                _path.push_back(intermediates[8]);
                break;
              }
          }
      }
    else
      {
        _type = shortestPath::pathType::AMBIG;
      }
    _path.push_back(codon2);
  }

  std::pair<double,shortestPath::pathType> 
  shortestPath::shortestPathImpl::process_path(const std::string &intermediate,
                                               const Sequence::GeneticCodes & code)
  {
    std::string tint = Sequence::Translate(intermediate.begin(),intermediate.end(),
                                           code);
    shortestPath::pathType t = shortestPath::pathType::AMBIG;
    double d = ( distances(t1[0],tint[0]) + distances(tint[0],t2[0]) );
    if ( (t1 != tint) && (tint != t2) )
      {
        t = shortestPath::pathType::NN;
      }
    else if ( ((t1 != tint) && (tint == t2)) ||
              ((t1 == tint) && (tint != t2)) )
      {
        t = shortestPath::pathType::SN;
      }
    else if ( t1==tint && tint==t2 )
      {
        t = shortestPath::pathType::SS;
      }
    return std::make_pair(d,t);
  }

  std::pair<double,shortestPath::pathType> 
  shortestPath::shortestPathImpl::process_path2(const std::string &intermediate1,
                                                const std::string &intermediate2,
                                                const Sequence::GeneticCodes & code)
  {
    std::string tint1 = Sequence::Translate(intermediate1.begin(),
                                            intermediate1.end(),code);
    std::string tint2 = Sequence::Translate(intermediate2.begin(),
                                            intermediate2.end(),code);
    shortestPath::pathType t = shortestPath::pathType::AMBIG;
    double d = (distances(t1[0],tint1[0]) + 
                distances(tint1[0],tint2[0]) + 
                distances(tint2[0],t2[0]));

    bool a = (t1==tint1),
      b = (tint1==tint2),
      c = (tint2==t2);

    if ( !a && !b && !c )
      {
        t = shortestPath::pathType::NNN;
      }
    else if ( (a && b && !c) ||
              (a && !b && c) ||
              (!a && b && c) )
      {
        t = shortestPath::pathType::SSN;
      }
    else if ( (!a && !b && c) ||
              (!a && b && !c) ||
              ( a && !b && !c) )
      {
        t = shortestPath::pathType::SNN;
      }

    else if (a && b && c)
      {
        t = shortestPath::pathType::SSS;
      }
    return std::make_pair(d,t);
  }
#endif

  shortestPath::shortestPath(const std::string &codon1,
                             const std::string &codon2,
                             const Sequence::GeneticCodes & code) 
    

  
  {
    try
      {
        impl = std::unique_ptr<shortestPath::shortestPathImpl>(new shortestPath::shortestPathImpl(codon1,codon2,code));
      }
    catch(std::runtime_error &e)
      {
        throw;
      }
    catch(std::exception &e)
      {
        throw (std::runtime_error(e.what()));
      }
    catch (...)
      {
        throw (std::runtime_error("caught exception of unknown type"));
      }
  }

  shortestPath::~shortestPath()
  {
  }

  shortestPath::pathType shortestPath::type() const
  { 
    return impl->_type;
  }

  shortestPath::const_iterator shortestPath::begin() const
  {
    return impl->_path.begin();
  }

  shortestPath::const_iterator shortestPath::end() const
  {
    return impl->_path.end();
  }

  double shortestPath::path_distance() const
  {
    return impl->_distance;
  }

  std::pair<unsigned,unsigned> mutsShortestPath(const std::string &codon1,
                                                const std::string &codon2, 
                                                const Sequence::GeneticCodes & code)
    

  
  {
    std::pair<unsigned,unsigned> rv;
    try
      {
        shortestPath sp(codon1,codon2,code);
        shortestPath::pathType type = sp.type();
        switch (type)
          {
          case shortestPath::pathType::S :
            {
              rv = std::make_pair(1,0);
              break;
            }
          case shortestPath::pathType::N :
            {
              rv = std::make_pair(0,1);
              break;
            }
          case shortestPath::pathType::SS :
            {
              rv = std::make_pair(2,0);
              break;
            }
          case shortestPath::pathType::SN :
            {
              rv = std::make_pair(1,1);
              break;
            }
          case shortestPath::pathType::NN :
            {
              rv = std::make_pair(0,2);
              break;
            }
          case shortestPath::pathType::SSS :
            {
              rv =  std::make_pair(3,0);
              break;
            }
          case shortestPath::pathType::SSN :
            {
              rv = std::make_pair(2,1);
              break;
            }
          case shortestPath::pathType::SNN :
            {
              rv = std::make_pair(1,2);
              break;
            }
          case shortestPath::pathType::NNN :
            {
              rv = std::make_pair(0,3);
              break;
            }
          case shortestPath::pathType::NONE :
            {
              rv = std::make_pair(0,0);
              break;
            }
          case shortestPath::pathType::AMBIG :
            {
              rv = std::make_pair(SEQMAXUNSIGNED,SEQMAXUNSIGNED);
              break;
            }
          }
        return std::make_pair(SEQMAXUNSIGNED,SEQMAXUNSIGNED);
      }
    catch (std::runtime_error &e)
      {
        throw;
      }
    return rv;
  }

  std::pair<unsigned,shortestPath::pathType> diffType(const std::string &codon1,
                                                      const std::string &codon2,
                                                      const Sequence::GeneticCodes & code)
    

  
  {
    try
      {
        shortestPath sp(codon1,codon2,code);
        shortestPath::pathType type = sp.type();
        //check preconditions
        if (type == shortestPath::pathType::AMBIG)
          {
            return std::make_pair(SEQMAXUNSIGNED,type);
          }
        unsigned site=0,ndiffs=0;
        for(unsigned i = 0 ; i < 3 ; ++i)
          {
            if (codon1[i] != codon2[i])
              {
                ++ndiffs;
                site = i;
              }
          }
        if( ndiffs != 1 )
          {
            return std::make_pair(SEQMAXUNSIGNED,type);
          }
        return std::make_pair(site,type);
      }
    catch (std::runtime_error &e)
      {
        throw;
      }
  }

  std::tuple<shortestPath::pathType,shortestPath::pathType,shortestPath::pathType>
  diffTypeMulti(const std::string &codon1,
                const std::string &codon2,
                const Sequence::GeneticCodes &code)
    

  
  {
    try
      {
        shortestPath::pathType p[3]; //one for each codon position
        std::string t1(codon1),t2(codon2);
        for(unsigned i = 0; i < 3 ; ++i)
          {
            //make codons that only differ at one position each
            std::swap(t1[i],t2[i]);
            std::pair<unsigned,shortestPath::pathType> dt1 = diffType(t1,codon1,code);
            std::pair<unsigned,shortestPath::pathType> dt2 = diffType(t2,codon2,code);
            if (dt1 == dt2)
              p[i] = dt1.second;
            else
              p[i] = shortestPath::pathType::AMBIG;
            //swap back...
            std::swap(t1[i],t2[i]);
          }
        return std::make_tuple(p[0],p[1],p[2]);
      }
    catch(std::runtime_error &e)
      {
        throw;
      }
  }
}