Merge branch 'master' of https://github.com/ms609/SlowQuartet

R/TreeManipulation.R

@@ -1,4 +1,4 @@
-#' Unshift Tree
+#' Add Tree to Start of List
 #' 
 #' Add a tree to the start of a list of trees
 #' 

README.md

@@ -1,8 +1,8 @@
-[![Project Status: Inactive.](http://www.repostatus.org/badges/latest/inactive.svg)](http://www.repostatus.org/#inactive)
+[![Project Status: Inactive.](https://www.repostatus.org/badges/latest/inactive.svg)](https://www.repostatus.org/#inactive)
 [![Build Status](https://travis-ci.org/ms609/Quartet.svg?branch=master)](https://travis-ci.org/ms609/Quartet)
 [![codecov](https://codecov.io/gh/ms609/Quartet/branch/master/graph/badge.svg)](https://codecov.io/gh/ms609/Quartet)
-[![CRAN Status Badge](http://www.r-pkg.org/badges/version/Quartet)](https://cran.r-project.org/package=Quartet)
-[![CRAN Downloads](http://cranlogs.r-pkg.org/badges/Quartet)](https://cran.r-project.org/package=Quartet)
+[![CRAN Status Badge](https://www.r-pkg.org/badges/version/Quartet)](https://cran.r-project.org/package=Quartet)
+[![CRAN Downloads](https://cranlogs.r-pkg.org/badges/Quartet)](https://cran.r-project.org/package=Quartet)
 [![DOI](https://zenodo.org/badge/80424189.svg)](https://zenodo.org/badge/latestdoi/80424189)
 
 # Quartet

cran-comments.md

@@ -2,25 +2,30 @@
 * local Windows 10 install, R 3.5.2
 * win-builder, with `check_win_devel()`
 * Ubuntu 14.04.5 LTS, R 3.4 and devel, via [Travis CI](https://travis-ci.org/ms609/Quartet)
-* R-hub, with `check_rhub()`
+* R-hub, with `check_rhub(platforms = platforms())`
 
   * Installation failed with PREPERROR on rhub/ubuntu-gcc-release and
-    rhub/fedora-clang-devel, on account of failures to install the required 
-    packages (`phangorn`; `git2r`).  This seems to be an issue outside
-    of my control; the Debian rhub installation reports success.
-
+    rhub/fedora-clang-devel (and other platforms), on account of failures 
+    to install a required dependency (`phangorn`).  
+    This seems to be an issue outside of my control.
+  * Installation failed with ERROR on ubuntu-gcc-devel and debian-gcc-patched:
+    > "there is no package called `lattice`"
+    The failure to install the dependency `lattice` seems to be outwith my 
+    control.
 
 ## R CMD check results
 There were no ERRORs or WARNINGs.
 
-There was 1 NOTE:
+There was one NOTE:
+
+ > Days since last update: 6
 
-* checking CRAN incoming feasibility ... NOTE
-  Maintainer: 'Smith Martin R. <[email protected]>'
-  
-  New submission
-  
-This package is a new submission.
+This release comes on the heels of v1.0.0 on 2019-01-16, which failed
+certain CRAN package checks -- Brian Ripley requested that these be fixed 
+immediately and kindly proposed some suggestions for how portability could be
+improved.
+After implementing these and testing the relevant platforms using `check_rhub`, 
+I believe that the issues have been resolved.
 
 ## Downstream dependencies
 There are currently no downstream dependencies for this package.

inst/REFERENCES.bib

@@ -1,3 +1,60 @@
+@article{Bluis2003,
+  author = {Bluis, John and Shin, Dong Guk},
+  doi = {10.1109/BIBE.2003.1188933},
+  journal = {Proceedings - 3rd IEEE Symposium on BioInformatics and BioEngineering, BIBE 2003},
+  pages = {87--94},
+  publisher = {IEEE},
+  title = {{Nodal distance algorithm: Calculating a phylogenetic tree comparison metric}},
+  year = {2003}
+}
+
+@incollection{Bocker2013,
+  address = {Berlin, Heidelberg},
+  author = {B{\"{o}}cker, Sebastian and Canzar, Stefan and Klau, Gunnar W.},
+  booktitle = {Algorithms in Bioinformatics. WABI 2013. Lecture Notes in Computer Science, vol 8126},
+  doi = {10.1007/978-3-642-40453-5_13},
+  editor = {Darling, A. and Stoye, J.},
+  pages = {156--169},
+  publisher = {Springer},
+  title = {{The generalized Robinson-Foulds metric}},
+  year = {2013}
+}
+
+
+@article{Bogdanowicz2012,
+  author = {Bogdanowicz, Damian and Giaro, Krzysztof},
+  doi = {10.1109/TCBB.2011.48},
+  issn = {15455963},
+  journal = {IEEE/ACM Transactions on Computational Biology and Bioinformatics},
+  number = {1},
+  pages = {150--160},
+  title = {{Matching split distance for unrooted binary phylogenetic trees}},
+  volume = {9},
+  year = {2012}
+}
+
+@article{Bogdanowicz2013,
+  author = {Bogdanowicz, Damian and Giaro, Krzysztof},
+  doi = {10.2478/amcs-2013-0050},
+  journal = {International Journal of Applied Mathematics and Computer Science},
+  number = {3},
+  pages = {669--684},
+  title = {{On a matching distance between rooted phylogenetic trees}},
+  volume = {23},
+  year = {2013}
+}
+
+@article{Bogdanowicz2017,
+  author = {Bogdanowicz, Damian and Giaro, Krzysztof},
+  doi = {10.1089/cmb.2016.0204},
+  journal = {Journal of Computational Biology},
+  number = {5},
+  pages = {422--435},
+  title = {{Comparing phylogenetic trees by matching nodes using the transfer distance between partitions}},
+  volume = {24},
+  year = {2017}
+}
+
 @article{Brodal2013,
   author = {Brodal, Gerth St{\o}lting and Fagerberg, Rolf and Mailund, Thomas and Pedersen, Christian N. S. and Sand, Andreas},
   doi = {10.1137/1.9781611973105.130},
@@ -61,6 +118,17 @@ @article{Marczewski1958
   year = {1958}
 }
 
+@article{Nye2006,
+  author = {Nye, Tom M W and Li{\`{o}}, Pietro and Gilks, Walter R.},
+  doi = {10.1093/bioinformatics/bti720},
+  journal = {Bioinformatics},
+  number = {1},
+  pages = {117--119},
+  title = {{A novel algorithm and web-based tool for comparing two alternative phylogenetic trees}},
+  volume = {22},
+  year = {2006}
+}
+
 @article{Penny1985,
   author = {Penny, David and Hendy, M. D.},
   doi = {10.2307/2413347},

src/int_stuff.h

@@ -3,8 +3,8 @@
   #define INT_STUFF_H
 
 #if __cplusplus <= 199711L
-  #define INTTYPE_N4 long long // fails -pedantic
-  #define INTTYPE_REST long long // fails -pedantic
+  #define INTTYPE_N4 long int // long long fails -pedantic
+  #define INTTYPE_REST long int // long long fails -pedantic
 #else 
   #include <cstdint>
   #define INTTYPE_N4 int_fast64_t // long long fails -pedantic

src/rQuartetDist.cpp

@@ -1,12 +1,9 @@
 #include <Rcpp.h>
 using namespace Rcpp;
-#include "int_stuff.h"
 
+#include "int_stuff.h"
 #include "QuartetDistanceCalculator.h"
 
-#include <R.h>
-#include <Rdefines.h>
-#include <Rinternals.h>
 #include <vector>
 
 //' Triplet and quartet distances with tqDist

src/rTripletDist.cpp

@@ -1,13 +1,9 @@
 #include <Rcpp.h>
 using namespace Rcpp;
-#include "int_stuff.h"
 
+#include "int_stuff.h"
 #include "TripletDistanceCalculator.h"
 
-#include <R.h>
-#include <Rdefines.h>
-#include <Rinternals.h>
-#include <cstdlib>
 #include <vector>
 
 //' @describeIn tqdist_QuartetDistance Triplet distance between two trees

vignettes/Tree-distance-metrics.Rmd

@@ -74,6 +74,8 @@ each pair of tips.
 The path difference metric [@Steel1993; `r Phangorn('path.dist')`] is the 
 square root of the sum of squares of each of these differences.
 
+The related nodal distance [@Bluis2003] is not considered further here.
+
 ## Partition metric
 The Robinson-Foulds (RF or 'partition') metric [@Robinson1981; @Steel1993;
 phangorn's `r Phangorn('RF.dist')` or this package's `RobinsonFoulds`]
@@ -372,5 +374,33 @@ all of the following desiderata:
  - Identifies pairs of trees that are more random than expected by chance
  - Does not reach its maximum value after relatively trivial rearrangements
 
+# Improvements to partition metrics
+
+The primary issue with the Robinson-Foulds distance is its conservative 
+pairing of clades: a grouping is either identical in both trees, or is 
+treated as different.  
+This makes no account for the possibility that one tree contains a grouping that
+is similar (though not quite identical) to a grouping in another.
+
+In order to incorporate some measure of similar-but-not-quite-identical clades,
+it is necessary to construct an optimal matching that pairs each clade in each
+tree with either a suitable clade in the other, or with no clade at all.
+
+Various approaches to this problem have been proposed; these typically search
+for a matching that minimises the total dissimilarity summed for all pairings.
+Approaches differ in how they penalize differences between paired groups.
+Bogdanowicz and Giaro [-@Bogdanowicz2012;-@Bogdanowicz2013;-@Bogdanowicz2017]
+calculate the size
+of the largest bipartition that is consistent with both input groups; 
+Nye _et al_. [-@Nye2006] normalize this number using the Jaccard index;
+Böcker _et al_. [-@Bocker2013] raise this value to an arbitrary exponent, and
+add a constraint that prohibits pairings that comprise contradictory groups.
+
+These approaches represent a refinement of the Robinson-Foulds metric,
+circumventing many of the issues raised above.  They are, however, more 
+computationally expensive, and lack an objective measure of clade similarity.
+I am not presently aware of an R implementation of any of these measures,
+though a related metric is planned for a future version of this package.
+
 # References