[Bioc-devel] PANTHER.db GOSLIM_ID hosed?

[Steve Lianoglou]

Thanks to all for tackling this so quickly.

-steve

On Mon, Nov 23, 2015 at 6:59 AM, Jim Hester <james.f.hester at gmail.com> wrote:
> Panther has been updated to the new 1.03, which should resolve this issue.
> Let us know if there are any further updates.
>
> On Fri, Nov 20, 2015 at 5:29 AM, Julius Müller <mail at jmuller.eu> wrote:
>
>> Dear Steve,
>>
>> thanks a lot for pointing out the error! The problem Martin described is
>> exactly the problem addressed in PANTHER.db version 1.02. However the
>> version available on bioconductor is not v1.02.
>> The method type2col in PANTHER.db 1.02 should look like this:
>>
>> type2col <- function(x){
>>
>>   column_names <-
>> c("family_id","goslim_id","ontology","uniprot_id","species","family_term","subfamily_term","class_id","class_term","go_id","go_term","component_go_id","component_term","evidence","evidence_type","confidence_code","entrez_id")
>>   names(column_names) <- .cols(x)
>>   column_names
>> }
>>
>>
>> What's puzzling to me is that there are unit tests included in the loading
>> routine, so checking the version currently available on bioconductor should
>> have choked executing PANTHER.db:::.test()
>> I remember when I first uploaded the new version to Jim, I have send an
>> old version by accident and then later on the correct version. This might
>> have caused some confusion.
>> Can somebody send me another aspera invite, then we can make sure that you
>> have got the right version?
>>
>> Many thanks and apologies if this caused any troubles,
>> Julius
>>
>>
>>
>> On 20/11/2015 09:46, Morgan, Martin wrote:
>>
>> From the head of
>>
>>
>> PANTHER.db:::.keys
>>
>> function (x, keytype)
>> {
>>     t2t <- .type2table(x)
>>     t2c <- .type2col(x)
>>
>> I wonder if many of the columns / keytypes are mislabeled, e.g.,
>>
>>
>> PANTHER.db:::.type2col(PANTHER.db)
>>
>>          CLASS_ID        CLASS_TERM      COMPONENT_ID    COMPONENT_TERM
>>       "family_id"       "goslim_id"        "ontology"      "uniprot_id"
>>   CONFIDENCE_CODE            ENTREZ          EVIDENCE     EVIDENCE_TYPE
>>         "species"     "family_term"  "subfamily_term"        "class_id"
>>         FAMILY_ID       FAMILY_TERM         GOSLIM_ID       GOSLIM_TERM
>>      "class_term"           "go_id"         "go_term" "component_go_id"
>>        PATHWAY_ID      PATHWAY_TERM           SPECIES    SUBFAMILY_TERM
>>  "component_term"        "evidence"   "evidence_type" "confidence_code"
>>           UNIPROT
>>       "entrez_id"
>>
>> probably because columns() is used in .type2* but was changed at some point to return sorted values
>>
>>
>> showMethods(PANTHER.db:::.type2col, includeDefs=TRUE)
>>
>> Function: .type2col (package PANTHER.db)
>> x="PANTHER.db"
>> function (x)
>> {
>>     cls <- c("family_id", "goslim_id", "ontology", "uniprot_id",
>>         "species", "family_term", "subfamily_term", "class_id",
>>         "class_term", "go_id", "go_term", "component_go_id",
>>         "component_term", "evidence", "evidence_type", "confidence_code",
>>         "entrez_id")
>>     names(cls) <- columns(x)
>>     cls
>> }
>>
>>
>> columns
>>
>> nonstandardGenericFunction for "columns" defined from package "AnnotationDbi"
>>
>> function (x)
>> {
>>     value <- standardGeneric("columns")
>>     sort(value)
>> }
>>
>>
>> selectMethod("columns", "PANTHER.db")
>>
>> Method Definition:
>>
>> function (x)
>> {
>>     c("FAMILY_ID", "GOSLIM_ID", "GOSLIM_TERM", "UNIPROT", "SPECIES",
>>         "FAMILY_TERM", "SUBFAMILY_TERM", "CLASS_ID", "CLASS_TERM",
>>         "PATHWAY_ID", "PATHWAY_TERM", "COMPONENT_ID", "COMPONENT_TERM",
>>         "EVIDENCE", "EVIDENCE_TYPE", "CONFIDENCE_CODE", "ENTREZ")
>> }
>>
>> Probably a better design is to define a package-global constant
>>
>> .TYPE2COLUMN = c(FAMILY_ID="family_id", GOSLIM_ID="goslim_id", ...)
>>
>> and implement columns,PANTHER.db-method as names(.TYPE2COLUMN)
>>
>> Martin Morgan
>> ________________________________________
>> From: Bioc-devel [bioc-devel-bounces at r-project.org] on behalf of Steve Lianoglou [lianoglou.steve at gene.com]
>> Sent: Friday, November 20, 2015 4:24 AM
>> To: bioc-devel at r-project.org
>> Subject: [Bioc-devel] PANTHER.db GOSLIM_ID hosed?
>>
>> Hi all,
>>
>> I'm running the latest Bioc + PANTHER.db
>>
>> I think we've got the GOSLIM_TERM where the GOSLIM_ID should be:
>>
>> In the vignette, we have this:
>>
>> R> go_ids <- head(keys(PANTHER.db,keytype="GOSLIM_ID"))
>> R> go_ids
>>
>> ## [1] "GO:0000003" "GO:0000165" "GO:0000166" "GO:0000228"
>> "GO:0000375" "GO:0000398"
>>
>> But what we really get is:
>>
>> R> head(keys(PANTHER.db,keytype="GOSLIM_ID"))
>>
>> [1] "2-arachidonoylglycerol biosynthesis"
>> [2] "5-Hydroxytryptamine biosynthesis"
>> [3] "5-Hydroxytryptamine degredation"
>> [4] "5HT1 type receptor mediated signaling pathway"
>>
>> R> packageVersion('PANTHER.db')
>> [1] '1.0.2'
>>
>> Thanks,
>> -steve
>>
>> _______________________________________________Bioc-devel at r-project.org mailing listhttps://stat.ethz.ch/mailman/listinfo/bioc-devel
>>
>> This email message may contain legally privileged and/or confidential information.  If you are not the intended recipient(s), or the employee or agent responsible for the delivery of this message to the intended recipient(s), you are hereby notified that any disclosure, copying, distribution, or use of this email message is prohibited.  If you have received this message in error, please notify the sender immediately by e-mail and delete this email message from your computer. Thank you.
>>
>>
>>
>> --
>> Julius Müller 1A All Saints Road OX3 7AU, Oxford United Kingdom Phone: +44
>> 750 828 0522 mail at jmuller.eu
>>
>>
>>
>
>         [[alternative HTML version deleted]]
>
> _______________________________________________
> Bioc-devel at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/bioc-devel



-- 
Steve Lianoglou
Computational Biologist
Genentech