gnu/packages/bioconductor.scm

   1 ;;; GNU Guix --- Functional package management for GNU
   2 ;;; Copyright © 2018 Ricardo Wurmus <rekado@elephly.net>
   3 ;;; Copyright © 2018 Roel Janssen <roel@gnu.org>
   4 ;;; Copyright © 2018 Tobias Geerinckx-Rice <me@tobias.gr>
   5 ;;;
   6 ;;; This file is part of GNU Guix.
   7 ;;;
   8 ;;; GNU Guix is free software; you can redistribute it and/or modify it
   9 ;;; under the terms of the GNU General Public License as published by
  10 ;;; the Free Software Foundation; either version 3 of the License, or (at
  11 ;;; your option) any later version.
  12 ;;;
  13 ;;; GNU Guix is distributed in the hope that it will be useful, but
  14 ;;; WITHOUT ANY WARRANTY; without even the implied warranty of
  15 ;;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  16 ;;; GNU General Public License for more details.
  17 ;;;
  18 ;;; You should have received a copy of the GNU General Public License
  19 ;;; along with GNU Guix.  If not, see <http://www.gnu.org/licenses/>.
  20
  21 (define-module (gnu packages bioconductor)
  22   #:use-module ((guix licenses) #:prefix license:)
  23   #:use-module (guix packages)
  24   #:use-module (guix download)
  25   #:use-module (guix build-system r)
  26   #:use-module (gnu packages)
  27   #:use-module (gnu packages cran)
  28   #:use-module (gnu packages compression)
  29   #:use-module (gnu packages statistics)
  30   #:use-module (gnu packages bioinformatics))
  31
  32 (define-public r-bsgenome-dmelanogaster-ucsc-dm6
  33   (package
  34     (name "r-bsgenome-dmelanogaster-ucsc-dm6")
  35     (version "1.4.1")
  36     (source (origin
  37               (method url-fetch)
  38               ;; We cannot use bioconductor-uri here because this tarball is
  39               ;; located under "data/annotation/" instead of "bioc/".
  40               (uri (string-append "https://www.bioconductor.org/packages/"
  41                                   "release/data/annotation/src/contrib/"
  42                                   "BSgenome.Dmelanogaster.UCSC.dm6_"
  43                                   version ".tar.gz"))
  44               (sha256
  45                (base32
  46                 "1bhj0rdgf7lspw4xby9y9mf7v7jxxz8001bc8vw8kf04rjsx6060"))))
  47     (properties
  48      `((upstream-name . "BSgenome.Dmelanogaster.UCSC.dm6")))
  49     (build-system r-build-system)
  50     ;; As this package provides little more than a very large data file it
  51     ;; doesn't make sense to build substitutes.
  52     (arguments `(#:substitutable? #f))
  53     (propagated-inputs
  54      `(("r-bsgenome" ,r-bsgenome)))
  55     (home-page
  56      "https://www.bioconductor.org/packages/BSgenome.Dmelanogaster.UCSC.dm6/")
  57     (synopsis "Full genome sequences for Fly")
  58     (description
  59      "This package provides full genome sequences for Drosophila
  60 melanogaster (Fly) as provided by UCSC (dm6) and stored in Biostrings
  61 objects.")
  62     (license license:artistic2.0)))
  63
  64 (define-public r-bsgenome-dmelanogaster-ucsc-dm3-masked
  65   (package
  66     (name "r-bsgenome-dmelanogaster-ucsc-dm3-masked")
  67     (version "1.3.99")
  68     (source (origin
  69               (method url-fetch)
  70               ;; We cannot use bioconductor-uri here because this tarball is
  71               ;; located under "data/annotation/" instead of "bioc/".
  72               (uri (string-append "http://www.bioconductor.org/packages/"
  73                                   "release/data/annotation/src/contrib/"
  74                                   "BSgenome.Dmelanogaster.UCSC.dm3.masked_"
  75                                   version ".tar.gz"))
  76               (sha256
  77                (base32
  78                 "1756csb09f1br9rj1l3f08qyh4hlymdbd0cfn8x3fq39dn45m5ap"))))
  79     (properties
  80      `((upstream-name . "BSgenome.Dmelanogaster.UCSC.dm3.masked")))
  81     (build-system r-build-system)
  82     (propagated-inputs
  83      `(("r-bsgenome" ,r-bsgenome)
  84        ("r-bsgenome-dmelanogaster-ucsc-dm3"
  85         ,r-bsgenome-dmelanogaster-ucsc-dm3)))
  86     (home-page "https://www.bioconductor.org/packages/BSgenome.Dmelanogaster.UCSC.dm3.masked/")
  87     (synopsis "Full masked genome sequences for Fly")
  88     (description
  89      "This package provides full masked genome sequences for Drosophila
  90 melanogaster (Fly) as provided by UCSC (dm3, April 2006) and stored in
  91 Biostrings objects.  The sequences are the same as in
  92 BSgenome.Dmelanogaster.UCSC.dm3, except that each of them has the 4 following
  93 masks on top: (1) the mask of assembly gaps (AGAPS mask), (2) the mask of
  94 intra-contig ambiguities (AMB mask), (3) the mask of repeats from
  95 RepeatMasker (RM mask), and (4) the mask of repeats from Tandem Repeats
  96 Finder (TRF mask).  Only the AGAPS and AMB masks are \"active\" by default.")
  97     (license license:artistic2.0)))
  98
  99 (define-public r-hpar
 100   (package
 101     (name "r-hpar")
 102     (version "1.22.2")
 103     (source
 104      (origin
 105        (method url-fetch)
 106        (uri (bioconductor-uri "hpar" version))
 107        (sha256
 108         (base32
 109          "1b72hvzasf6q739gmx6jblbzzyq22l7crrkbbfkihv3v7s94g388"))))
 110     (build-system r-build-system)
 111     (home-page "https://bioconductor.org/packages/hpar/")
 112     (synopsis "Human Protein Atlas in R")
 113     (description "This package provides a simple interface to and data from
 114 the Human Protein Atlas project.")
 115     (license license:artistic2.0)))
 116
 117 (define-public r-regioner
 118   (package
 119     (name "r-regioner")
 120     (version "1.12.0")
 121     (source
 122      (origin
 123        (method url-fetch)
 124        (uri (bioconductor-uri "regioneR" version))
 125        (sha256
 126         (base32
 127          "09bzlaqdgy7wmzly3zc9y2da50d07mlixlnpaxdxpiwdk8qmhxsb"))))
 128     (properties `((upstream-name . "regioneR")))
 129     (build-system r-build-system)
 130     (propagated-inputs
 131      `(("r-memoise" ,r-memoise)
 132        ("r-genomicranges" ,r-genomicranges)
 133        ("r-bsgenome" ,r-bsgenome)
 134        ("r-rtracklayer" ,r-rtracklayer)
 135        ("r-genomeinfodb" ,r-genomeinfodb)
 136        ("r-iranges" ,r-iranges)
 137        ("r-s4vectors" ,r-s4vectors)))
 138     (home-page "https://bioconductor.org/packages/regioneR/")
 139     (synopsis "Association analysis of genomic regions")
 140     (description "This package offers a statistical framework based on
 141 customizable permutation tests to assess the association between genomic
 142 region sets and other genomic features.")
 143     (license license:artistic2.0)))
 144
 145 (define-public r-diffbind
 146   (package
 147     (name "r-diffbind")
 148     (version "2.8.0")
 149     (source
 150      (origin
 151        (method url-fetch)
 152        (uri (bioconductor-uri "DiffBind" version))
 153        (sha256
 154         (base32
 155          "1w1hybzd732ccg3q8zhirwfilq8sx3frv1x98zfyj3svzw98fish"))))
 156     (properties `((upstream-name . "DiffBind")))
 157     (build-system r-build-system)
 158     (inputs
 159      `(("zlib" ,zlib)))
 160     (propagated-inputs
 161      `(("r-amap" ,r-amap)
 162        ("r-biocparallel" ,r-biocparallel)
 163        ("r-deseq2" ,r-deseq2)
 164        ("r-dplyr" ,r-dplyr)
 165        ("r-edger" ,r-edger)
 166        ("r-genomicalignments" ,r-genomicalignments)
 167        ("r-genomicranges" ,r-genomicranges)
 168        ("r-ggplot2" ,r-ggplot2)
 169        ("r-ggrepel" ,r-ggrepel)
 170        ("r-gplots" ,r-gplots)
 171        ("r-iranges" ,r-iranges)
 172        ("r-lattice" ,r-lattice)
 173        ("r-limma" ,r-limma)
 174        ("r-locfit" ,r-locfit)
 175        ("r-rcolorbrewer" , r-rcolorbrewer)
 176        ("r-rcpp" ,r-rcpp)
 177        ("r-rsamtools" ,r-rsamtools)
 178        ("r-s4vectors" ,r-s4vectors)
 179        ("r-summarizedexperiment" ,r-summarizedexperiment)
 180        ("r-systempiper" ,r-systempiper)
 181        ("r-zlibbioc" ,r-zlibbioc)))
 182     (home-page "http://bioconductor.org/packages/DiffBind")
 183     (synopsis "Differential binding analysis of ChIP-Seq peak data")
 184     (description
 185      "This package computes differentially bound sites from multiple
 186 ChIP-seq experiments using affinity (quantitative) data.  Also enables
 187 occupancy (overlap) analysis and plotting functions.")
 188     (license license:artistic2.0)))
 189
 190 (define-public r-ripseeker
 191   (package
 192     (name "r-ripseeker")
 193     (version "1.20.0")
 194     (source
 195      (origin
 196        (method url-fetch)
 197        (uri (bioconductor-uri "RIPSeeker" version))
 198        (sha256
 199         (base32
 200          "0y9cvzqslfxj3z9mnp47mknff0pky2g5x8x1z1s5yjcx35q89xfi"))))
 201     (properties `((upstream-name . "RIPSeeker")))
 202     (build-system r-build-system)
 203     (propagated-inputs
 204      `(("r-s4vectors" ,r-s4vectors)
 205        ("r-iranges" ,r-iranges)
 206        ("r-genomicranges" ,r-genomicranges)
 207        ("r-summarizedexperiment" ,r-summarizedexperiment)
 208        ("r-rsamtools" ,r-rsamtools)
 209        ("r-genomicalignments" ,r-genomicalignments)
 210        ("r-rtracklayer" ,r-rtracklayer)))
 211     (home-page "http://bioconductor.org/packages/RIPSeeker")
 212     (synopsis
 213      "Identifying protein-associated transcripts from RIP-seq experiments")
 214     (description
 215      "This package infers and discriminates RIP peaks from RIP-seq alignments
 216 using two-state HMM with negative binomial emission probability.  While
 217 RIPSeeker is specifically tailored for RIP-seq data analysis, it also provides
 218 a suite of bioinformatics tools integrated within this self-contained software
 219 package comprehensively addressing issues ranging from post-alignments
 220 processing to visualization and annotation.")
 221     (license license:gpl2)))
 222
 223 (define-public r-multtest
 224   (package
 225     (name "r-multtest")
 226     (version "2.36.0")
 227     (source
 228      (origin
 229        (method url-fetch)
 230        (uri (bioconductor-uri "multtest" version))
 231        (sha256
 232         (base32
 233          "11949h2kglw13x8haaj4clg4jim1mwh5n98n9zxp9mmgn01z1lp0"))))
 234     (build-system r-build-system)
 235     (propagated-inputs
 236      `(("r-survival" ,r-survival)
 237        ("r-biocgenerics" ,r-biocgenerics)
 238        ("r-biobase" ,r-biobase)
 239        ("r-mass" ,r-mass)))
 240     (home-page "http://bioconductor.org/packages/multtest")
 241     (synopsis "Resampling-based multiple hypothesis testing")
 242     (description
 243      "This package can do non-parametric bootstrap and permutation
 244 resampling-based multiple testing procedures (including empirical Bayes
 245 methods) for controlling the family-wise error rate (FWER), generalized
 246 family-wise error rate (gFWER), tail probability of the proportion of
 247 false positives (TPPFP), and false discovery rate (FDR).  Several choices
 248 of bootstrap-based null distribution are implemented (centered, centered
 249 and scaled, quantile-transformed).  Single-step and step-wise methods are
 250 available.  Tests based on a variety of T- and F-statistics (including
 251 T-statistics based on regression parameters from linear and survival models
 252 as well as those based on correlation parameters) are included.  When probing
 253 hypotheses with T-statistics, users may also select a potentially faster null
 254 distribution which is multivariate normal with mean zero and variance
 255 covariance matrix derived from the vector influence function.  Results are
 256 reported in terms of adjusted P-values, confidence regions and test statistic
 257 cutoffs.  The procedures are directly applicable to identifying differentially
 258 expressed genes in DNA microarray experiments.")
 259     (license license:lgpl3)))
 260
 261 (define-public r-chippeakanno
 262   (package
 263     (name "r-chippeakanno")
 264     (version "3.14.0")
 265     (source
 266      (origin
 267        (method url-fetch)
 268        (uri (bioconductor-uri "ChIPpeakAnno" version))
 269        (sha256
 270         (base32
 271          "1kcnc3cnmrhdk1x7q3y6zsz09pgd3xn9xy1hfbxz48cajlb18ad0"))))
 272     (properties `((upstream-name . "ChIPpeakAnno")))
 273     (build-system r-build-system)
 274     (propagated-inputs
 275      `(("r-biocgenerics" ,r-biocgenerics)
 276        ("r-biostrings" ,r-biostrings)
 277        ("r-delayedarray" ,r-delayedarray)
 278        ("r-go-db" ,r-go-db)
 279        ("r-biomart" ,r-biomart)
 280        ("r-bsgenome" ,r-bsgenome)
 281        ("r-genomicfeatures" ,r-genomicfeatures)
 282        ("r-genomicranges" ,r-genomicranges)
 283        ("r-genomeinfodb" ,r-genomeinfodb)
 284        ("r-iranges" ,r-iranges)
 285        ("r-matrixstats" ,r-matrixstats)
 286        ("r-annotationdbi" ,r-annotationdbi)
 287        ("r-limma" ,r-limma)
 288        ("r-multtest" ,r-multtest)
 289        ("r-rbgl" ,r-rbgl)
 290        ("r-graph" ,r-graph)
 291        ("r-biocinstaller" ,r-biocinstaller)
 292        ("r-regioner" ,r-regioner)
 293        ("r-dbi" ,r-dbi)
 294        ("r-ensembldb" ,r-ensembldb)
 295        ("r-biobase" ,r-biobase)
 296        ("r-s4vectors" ,r-s4vectors)
 297        ("r-seqinr" ,r-seqinr)
 298        ("r-idr" ,r-idr)
 299        ("r-genomicalignments" ,r-genomicalignments)
 300        ("r-summarizedexperiment" ,r-summarizedexperiment)
 301        ("r-rsamtools" ,r-rsamtools)
 302        ("r-venndiagram" ,r-venndiagram)))
 303     (home-page "http://bioconductor.org/packages/ChIPpeakAnno")
 304     (synopsis "Peaks annotation from ChIP-seq and ChIP-chip experiments")
 305     (description
 306      "The package includes functions to retrieve the sequences around the peak,
 307 obtain enriched Gene Ontology (GO) terms, find the nearest gene, exon, miRNA or
 308 custom features such as most conserved elements and other transcription factor
 309 binding sites supplied by users.  Starting 2.0.5, new functions have been added
 310 for finding the peaks with bi-directional promoters with summary statistics
 311 (peaksNearBDP), for summarizing the occurrence of motifs in peaks
 312 (summarizePatternInPeaks) and for adding other IDs to annotated peaks or
 313 enrichedGO (addGeneIDs).")
 314     (license license:gpl2+)))
 315
 316 (define-public r-marray
 317   (package
 318     (name "r-marray")
 319     (version "1.58.0")
 320     (source (origin
 321               (method url-fetch)
 322               (uri (bioconductor-uri "marray" version))
 323               (sha256
 324                (base32 "0539flh3y1qy5b1bamkfwbskis765c5s33v1y9j51n33mxb9h08d"))))
 325     (build-system r-build-system)
 326     (propagated-inputs
 327      `(("r-limma" ,r-limma)))
 328     (home-page "http://bioconductor.org/packages/marray")
 329     (synopsis "Exploratory analysis for two-color spotted microarray data")
 330     (description "This package contains class definitions for two-color spotted
 331 microarray data.  It also includes fuctions for data input, diagnostic plots,
 332 normalization and quality checking.")
 333     (license license:lgpl2.0+)))
 334
 335 (define-public r-cghbase
 336   (package
 337    (name "r-cghbase")
 338    (version "1.40.0")
 339    (source (origin
 340             (method url-fetch)
 341             (uri (bioconductor-uri "CGHbase" version))
 342             (sha256
 343              (base32 "1hf44vma3kgwr61kjbszvfxkava8bjqnam1mdncqvczbypb2xwaq"))))
 344    (properties `((upstream-name . "CGHbase")))
 345    (build-system r-build-system)
 346    (propagated-inputs
 347     `(("r-biobase" ,r-biobase)
 348       ("r-marray" ,r-marray)))
 349    (home-page "http://bioconductor.org/packages/CGHbase")
 350    (synopsis "Base functions and classes for arrayCGH data analysis")
 351    (description "This package contains functions and classes that are needed by
 352 the @code{arrayCGH} packages.")
 353    (license license:gpl2+)))
 354
 355 (define-public r-cghcall
 356   (package
 357    (name "r-cghcall")
 358    (version "2.42.0")
 359    (source (origin
 360             (method url-fetch)
 361             (uri (bioconductor-uri "CGHcall" version))
 362             (sha256
 363              (base32 "0y71vfxv9x0am3xvv520yr95cb7m7y92dhdx1vkqki80jrmf12dz"))))
 364    (properties `((upstream-name . "CGHcall")))
 365    (build-system r-build-system)
 366    (propagated-inputs
 367     `(("r-biobase" ,r-biobase)
 368       ("r-cghbase" ,r-cghbase)
 369       ("r-impute" ,r-impute)
 370       ("r-dnacopy" ,r-dnacopy)
 371       ("r-snowfall" ,r-snowfall)))
 372    (home-page "http://bioconductor.org/packages/CGHcall")
 373    (synopsis "Base functions and classes for arrayCGH data analysis")
 374    (description "This package contains functions and classes that are needed by
 375 @code{arrayCGH} packages.")
 376    (license license:gpl2+)))
 377
 378 (define-public r-qdnaseq
 379   (package
 380     (name "r-qdnaseq")
 381     (version "1.16.0")
 382     (source (origin
 383               (method url-fetch)
 384               (uri (bioconductor-uri "QDNAseq" version))
 385               (sha256
 386                (base32 "1pj69mfyxwfd0d7h4kls9xq96sdc55y3rv20qpla50hw9libcwwd"))))
 387     (properties `((upstream-name . "QDNAseq")))
 388     (build-system r-build-system)
 389     (propagated-inputs
 390      `(("r-biobase" ,r-biobase)
 391        ("r-biocparallel" ,r-biocparallel)
 392        ("r-cghbase" ,r-cghbase)
 393        ("r-cghcall" ,r-cghcall)
 394        ("r-dnacopy" ,r-dnacopy)
 395        ("r-genomicranges" ,r-genomicranges)
 396        ("r-iranges" ,r-iranges)
 397        ("r-matrixstats" ,r-matrixstats)
 398        ("r-r-utils" ,r-r-utils)
 399        ("r-rsamtools" ,r-rsamtools)))
 400     (home-page "http://bioconductor.org/packages/QDNAseq")
 401     (synopsis "Quantitative DNA sequencing for chromosomal aberrations")
 402     (description "The genome is divided into non-overlapping fixed-sized bins,
 403 number of sequence reads in each counted, adjusted with a simultaneous
 404 two-dimensional loess correction for sequence mappability and GC content, and
 405 filtered to remove spurious regions in the genome.  Downstream steps of
 406 segmentation and calling are also implemented via packages DNAcopy and CGHcall,
 407 respectively.")
 408     (license license:gpl2+)))
 409
 410 (define-public r-bayseq
 411   (package
 412     (name "r-bayseq")
 413     (version "2.14.0")
 414     (source
 415      (origin
 416        (method url-fetch)
 417        (uri (bioconductor-uri "baySeq" version))
 418        (sha256
 419         (base32
 420          "0hbmm01a8libara9mbxknpk0wzarwfngnfwlmhpww91a0cmy5klg"))))
 421     (properties `((upstream-name . "baySeq")))
 422     (build-system r-build-system)
 423     (propagated-inputs
 424      `(("r-abind" ,r-abind)
 425        ("r-edger" ,r-edger)
 426        ("r-genomicranges" ,r-genomicranges)))
 427     (home-page "https://bioconductor.org/packages/baySeq/")
 428     (synopsis "Bayesian analysis of differential expression patterns in count data")
 429     (description
 430      "This package identifies differential expression in high-throughput count
 431 data, such as that derived from next-generation sequencing machines,
 432 calculating estimated posterior likelihoods of differential expression (or
 433 more complex hypotheses) via empirical Bayesian methods.")
 434     (license license:gpl3)))