getChromInfoFromEnsembl(species,
                        release=NA, division=NA, use.grch37=FALSE,
                        assembled.molecules.only=FALSE,
                        include.non_ref.sequences=FALSE,
                        include.contigs=FALSE,
                        include.clones=FALSE,
                        map.NCBI=FALSE,
                        recache=FALSE,
                        as.Seqinfo=FALSE)

## ---------------------------------------------------------------------
## A. BASIC EXAMPLES
## ---------------------------------------------------------------------

## Internet access required!

## === Worm ===
## https://uswest.ensembl.org/Caenorhabditis_elegans

celegans <- getChromInfoFromEnsembl("celegans")
attr(celegans, "species_info")

getChromInfoFromEnsembl("celegans", as.Seqinfo=TRUE)

celegans <- getChromInfoFromEnsembl("celegans", map.NCBI=TRUE)

## === Yeast ===
## https://uswest.ensembl.org/Saccharomyces_cerevisiae

scerevisiae <- getChromInfoFromEnsembl("scerevisiae")
attr(scerevisiae, "species_info")

getChromInfoFromEnsembl("scerevisiae", as.Seqinfo=TRUE)

scerevisiae <- getChromInfoFromEnsembl("scerevisiae", map.NCBI=TRUE)

## Arabidopsis thaliana:
athaliana <- getChromInfoFromEnsembl("athaliana", division="plants",
                                     map.NCBI=TRUE)
attr(athaliana, "species_info")

## ---------------------------------------------------------------------
## Temporary stuff that needs to go away...
## ---------------------------------------------------------------------

## TODO: Check all species for which an NCBI assembly is registered!
## Checked so far (with current Ensembl release i.e. 99):
## - celegans       OK
## - scerevisiae    OK
## - athaliana      OK
## - btaurus        OK
## - sscrofa        OK

## Not run: 
## WORK IN PROGRESS!!!
library(GenomeInfoDb)

.do_join <- GenomeInfoDb:::.do_join
.map_Ensembl_seqlevels_to_NCBI_seqlevels <-
    GenomeInfoDb:::.map_Ensembl_seqlevels_to_NCBI_seqlevels

.map_Ensembl_seqlevels_to_NCBI_seqlevels(
    paste0("ENS_", 1:26),
    CharacterList(c(list(c(aa="INSDC1", bb="GNBK7"), c("INSDC2", "RefSeq3")),
                    rep(list(NULL), 23), list("NCBI_7"))),
    paste0("NCBI_", 1:10),
    paste0("GNBK", c(1:8, NA, 9)),
    c(paste0("REFSEQ", c(1:7, 1, 1)), NA),
    verbose=TRUE
)

map_to_NCBI <- function(Ensembl_chrom_info, NCBI_chrom_info,
                        special_mappings=NULL)
{
    .map_Ensembl_seqlevels_to_NCBI_seqlevels(
         Ensembl_chrom_info[ , "name"],
         Ensembl_chrom_info[ , "synonyms"],
         NCBI_chrom_info[ , "SequenceName"],
         NCBI_chrom_info[ , "GenBankAccn"],
         NCBI_chrom_info[ , "RefSeqAccn"],
         special_mappings=special_mappings,
         verbose=TRUE)
}

## ------------------------------------------------------------------------
## Human
## https://uswest.ensembl.org/Homo_sapiens/
## Based on GRCh38.p13 (GCA_000001405.28)

## Return 944 rows
human_chrom_info <- getChromInfoFromEnsembl("hsapiens")
#                 1 id: 131550  <- ref chromosome
# CHR_HSCHR1_1_CTG3 id: 131561  <- non-ref chromosome
#     HSCHR1_1_CTG3 id: 131562  <- scaffold (no scaffold is non_ref)

## Map to NCBI
## Summary:
## - 639/640 NCBI sequences are reverse-mapped.
## - Restricted mapping is one-to-one.
GRCh38.p13 <- getChromInfoFromNCBI("GRCh38.p13")
L2R <- map_to_NCBI(human_chrom_info, GRCh38.p13)
## The only sequence in GRCh38.p13 that cannot be mapped to Ensembl is
## HG2139_PATCH (was introduced in GRCh38.p2)! Why? What's special about
## this patch?
GRCh38.p13$mapped <- tabulate(L2R, nbins=nrow(GRCh38.p13)) != 0L
table(GRCh38.p13$SequenceRole, GRCh38.p13$mapped)
#                        FALSE TRUE
#   assembled-molecule       0   25
#   alt-scaffold             0  261
#   unlocalized-scaffold     0   42
#   unplaced-scaffold        0  127
#   pseudo-scaffold          0    0
#   fix-patch                1  112
#   novel-patch              0   72
human_chrom_info <- .do_join(human_chrom_info, GRCh38.p13, L2R)
table(human_chrom_info$SequenceRole, human_chrom_info$toplevel)
#                       FALSE TRUE
#  assembled-molecule       0   25
#  alt-scaffold           261    0
#  unlocalized-scaffold     0   42
#  unplaced-scaffold        0  127
#  pseudo-scaffold          0    0
#  fix-patch              112    0
#  novel-patch             72    0

#hsa_seqlevels <- readRDS("hsapiens_gene_ensembl_txdb_seqlevels.rds")

## ------------------------------------------------------------------------
## Mouse
## https://uswest.ensembl.org/Mus_musculus/
## Based on GRCm38.p6 (GCA_000001635.8)

## Return 258 rows
mouse_chrom_info <- getChromInfoFromEnsembl("mmusculus")

## Map to NCBI
## Summary:
## - 139/239 NCBI sequences are reverse-mapped.
## - Restricted mapping is NOT one-to-one: 2 Ensembl sequences (NC_005089.1
##   and MT) are both mapped to NCBI MT.
GRCm38.p6 <- getChromInfoFromNCBI("GRCm38.p6")
L2R <- map_to_NCBI(mouse_chrom_info, GRCm38.p6)
## 100 sequences in GRCm38.p6 are not mapped:
GRCm38.p6$mapped <- tabulate(L2R, nbins=nrow(GRCm38.p6)) != 0L
table(GRCm38.p6$SequenceRole, GRCm38.p6$mapped)
#                        FALSE TRUE
#   assembled-molecule       0   22
#   alt-scaffold            99    0
#   unlocalized-scaffold     0   22
#   unplaced-scaffold        0   22
#   pseudo-scaffold          0    0
#   fix-patch                1   64
#   novel-patch              0    9
## OK so Ensembl doesn't include the alt-scaffolds for Mouse. BUT WHAT
## HAPPENED TO THIS ONE fix-patch SEQUENCE (MG4237_PATCH) THAT IS NOT
## MAPPED? Found it in seq_region_synonym table! It's seq_region_id=100405.
## Hey but that seq_region_id is **NOT** in the seq_region table!!! THIS
## VIOLATES FOREIGN KEY CONSTRAINT!!!!
mouse_chrom_info <- .do_join(mouse_chrom_info, GRCm38.p6, L2R)
## Ensembl does NOT comsider NC_005089.1 (duplicate entry for MT) toplevel:
mouse_chrom_info[mouse_chrom_info$SequenceName 
#            name length coord_system                      synonyms toplevel
# 184 NC_005089.1  16299     scaffold                                  FALSE
# 201          MT  16299   chromosome NC_005089.1, chrM, AY172335.1     TRUE
#     SequenceName GenBankAccn  RefSeqAccn
# 184           MT  AY172335.1 NC_005089.1
# 201           MT  AY172335.1 NC_005089.1

## ------------------------------------------------------------------------
## Rat
## https://uswest.ensembl.org/Rattus_norvegicus/
## Based on Rnor_6.0 (GCA_000001895.4)

# Return 1418 rows
rat_chrom_info <- getChromInfoFromEnsembl("rnorvegicus")

## Map to NCBI
## Summary:
## - 955/955 NCBI sequences are reverse-mapped.
## - Reverse mapping is one-to-many: 2 Ensembl sequences (NC_001665.2 and MT)
##   are mapped to NCBI MT.
Rnor_6.0 <- getChromInfoFromNCBI("Rnor_6.0")
L2R <- map_to_NCBI(rat_chrom_info, Rnor_6.0)
rat_chrom_info <- .do_join(rat_chrom_info, Rnor_6.0, L2R)

## Ensembl does NOT comsider NC_001665.2 (duplicate entry for MT) toplevel:
rat_chrom_info[rat_chrom_info$SequenceName 
#             name length coord_system                      synonyms toplevel
# 1417 NC_001665.2  16313     scaffold                                  FALSE
# 1418          MT  16313   chromosome NC_001665.2, AY172581.1, chrM     TRUE
#      SequenceName GenBankAccn  RefSeqAccn
# 1417           MT  AY172581.1 NC_001665.2
# 1418           MT  AY172581.1 NC_001665.2

table(rat_chrom_info$SequenceRole, rat_chrom_info$toplevel)
#                        FALSE TRUE
#   assembled-molecule       1   23
#   alt-scaffold             0    0
#   unlocalized-scaffold     0  354
#   unplaced-scaffold        0  578
#   pseudo-scaffold          0    0
#   fix-patch                0    0
#   novel-patch              0    0

## End(Not run)