[R] Using apply function on duplicates in a data.frame

[David Winsemius]

On Jan 31, 2010, at 6:05 PM, Sunny Srivastava wrote:

> Dear R-Helpers,
> I have a data.frame (df) and the head of data.frame looks like
>
>     ProbeUID ControlType     ProbeName GeneName SystematicName
> 1665     1577           0 pSysX_50_22_1 pSysX_50       pSysX_50
> 5422     5147           0  pSysX_49_8_1 pSysX_49       pSysX_49
> 4042     3843           0 pSysX_51_18_1 pSysX_51       pSysX_51
> 3646     3466           0   sll1514_0_2  sll1514        sll1514
> 2946     2807           0   sll1514_0_1  sll1514        sll1514
> 624       582           0  pSysX_49_8_2 pSysX_49       pSysX_49
>
>     Description   logFC AveExpr       t    P.Value  adj.P.Val
> 1665     Unknown  4.3887  9.5662  61.038 1.0938e-08 9.4449e-05
> 5422     Unknown -3.5251  6.9103 -35.908 1.7596e-07 3.5912e-04
> 4042     Unknown  2.5302  8.7497  35.112 1.9786e-07 3.5912e-04
> 3646     Unknown  2.3457 11.1678  33.962 2.3549e-07 3.5912e-04
> 2946     Unknown  2.3151 11.3153  32.689 2.8751e-07 3.5912e-04
> 624      Unknown -3.6256  6.8986 -31.777 3.3333e-07 3.5912e-04
>          B
> 1665 9.8342
> 5422 8.1650
> 4042 8.0758
> 3646 7.9408
> 2946 7.7822
> 624  7.6622
>

tdf <- tapply(df$logFC, df$GeneName, mean)
ndf <- dataframe(Gnames = names(tdf), mn.logFC= tdf)

> I want to "collapse" this data frame into a new data.frame so that the
> df$GeneName contains no duplicate GeneNames (for eg: sll1514) AND the
> df$logFC contains the average of df$logFC corresponding to these  
> GeneNames
> (which had duplicate genenames).
>
> I am aware of an inefficient strategy using loops, but I believe  
> that there
> should be a way using Apply functions or may be plyr?
>
> I am not able to think of one at the moment.  Can you please help me?

David Winsemius, MD
Heritage Laboratories
West Hartford, CT