[BioC] GDS2eSet problems

Sean Davis sdavis2 at mail.nih.gov
Thu Aug 9 02:43:51 CEST 2007


Martin Morgan wrote:
> Wyatt --
>
> Not that I have a solution, but if I
>
>   
>> debug(GDS2eSet)
>> GDS2eSet(gds961)
>>     
>
> and then step through the function (pressing the 'n' key) until just
> before the line "new('ExpressionSet', ...)", and then take a look at
> the row names of the expression data and of the feature data, I find
>
> Browse[1]> which(rownames(expr) != rownames(pData(featuredata)))
>  [1] 12626 12627 12628 12629 12630 12631 12632 12633 12634 12635 12636 12637
> [13] 12638 12639 12640 12641 12642 12643 12644 12645 12646 12647 12648
> Browse[1]> rownames(expr)[bad]
>  [1] "788_s_at"  "119_at"    "1215_at"   "1216_at"   "124_i_at"  "125_r_at" 
>  [7] "127_at"    "1301_s_at" "1302_s_at" "132_at"    "1429_at"   "1502_s_at"
> [13] "1829_at"   "1864_at"   "1889_s_at" "1982_s_at" "36969_at"  "383_at"   
> [19] "397_at"    "412_s_at"  "426_at"    "439_at"    "787_at"   
> Browse[1]> rownames(pData(featuredata))[bad]
>  [1] "119_at"    "1215_at"   "1216_at"   "124_i_at"  "125_r_at"  "127_at"   
>  [7] "1301_s_at" "1302_s_at" "132_at"    "1429_at"   "1502_s_at" "1829_at"  
> [13] "1864_at"   "1889_s_at" "1982_s_at" "36969_at"  "383_at"    "397_at"   
> [19] "412_s_at"  "426_at"    "439_at"    "787_at"    "788_s_at" 
>
> Position 12626 seems to be excised and re-inserted at 12648. So the
> problem is somewhere before the creation of ExpressionSet -- it could
> be that GDS2eSet needs to sort the rows of featureData to match the
> order of exprs, or that there is actually an error in the data coming
> from GEO. The package maintainer (copied on this email; use
> packageDescription to get this information) will likely chime in.
>   

Thanks, Martin, for doing the debugging.  This was a sorting problem and 
is now fixed in the Release branch (2.0) and in the development 
repository. 

Wyatt, once a new build is completed, the update should fix the issue.  
Let me know if you need it sooner and I can send you the fix directly.

Sean



More information about the Bioconductor mailing list