R help - May 2012 - help with melt/cast in reshape-package

I'm sorry everyone for the inconvenience of spamming the R-help...

Here's the complete post:


Hi everyone,
>
> Since it's quite a while that I used the reshape package, I now feel
kind
> of rusty.
>
> I have a data.frame like this:
>
>
>
> id          Sample.Name               Marker           Allele.1
> Allele.2            sample_id            species
> 1            01_primer01                  Dalb01           165
>      179                  SH233             D. madagascariensis
> 2            01_primer04                  Dalb04            221
>       225                  SH233             D. madagascariensis
> 3            01_primer08                  Dalb08            218
>       218                  SH233             D. madagascariensis
> 4            01_primer10                  Dalb10            134
>       134                  SH233             D. madagascariensis
> 5            01_primer14                  Dalb14            250
>       250                  SH233             D. madagascariensis
> 6            01_primer16                  Dalb16            232
>       232                  SH233             D. madagascariensis
>
> this was just the head(), in fact, the sample_id col has in fact different
> ids, I would like to aggregate them into one
> and I would like to get something like this:
>
> species                           sample_id             Marker1_Allele1
>        Marker1_Allele2          Marker2_Allele1          Marker2_Allele2
>      Marker3_Allele1         Marker3_Allele2           etc. (with 35
> different markers)
>
> D. madagascariensis       SH233                 179
>               225                                  134
>      244
>
> I tried to prepare the cast() but didn't quite figure out how to
achieve
> this. I wanted to create a col with the MarkerX_AlleleY 's, but
didn't find
> any useful solutions. The sample_id's should be aggregated so that each
> sample_id needs just one row. Should I  just merge columns Allele.1,
> Allele.2 and sample_id?
>
I'm kind of stuck, but would appreciate any help on the columns to be
merged.

Thanks a lot
Martin Schilling


>
>
	[[alternative HTML version deleted]]

Hi everyone,

I have an issue with the reshape package. Since it's quite a while that I
used it, I feel kind of rusty...

I got a data.frame like this


>
>>  id          Sample.Name               Marker           Allele.1
>>   Allele.2            sample_id            species
>> 1            01_primer01                  Dalb01           165
>>        179                  SH233             D. madagascariensis
>> 2            01_primer04                  Dalb04            221
>>       225                  SH233             D. madagascariensis
>> 3            01_primer08                  Dalb08            218
>>       218                  SH233             D. madagascariensis
>> 4            01_primer10                  Dalb10            134
>>       134                  SH233             D. madagascariensis
>> 5            01_primer14                  Dalb14            250
>>       250                  SH233             D. madagascariensis
>> 6            01_primer16                  Dalb16            232
>>       232                  SH233             D. madagascariensis
>>
>> this was just the head(), in fact, the sample_id col has different ids,
I
>> would like to aggregate them into one
>> and I would like to get something like this:
>>
>> species                           sample_id             Marker1_Allele1
>>        Marker1_Allele2          Marker2_Allele1         
Marker2_Allele2
>>      (etc. 35 markers)
>>
>> D. madagascariensis       SH233                 179
>>               225                                  134
>>      244                              ......
>>
>
I tried to prepare the cast() but didn't quite figure out how to achieve
this.  I tried to first merge with the following:

                genMelt <- melt(geno, id = c(1:2, 5:6))

then I created a column:

               genMelt$Locus <- substring(as.character(genMelt$Panel),5, 6)

               genMelt$Locus <- paste(genMelt$Locus, genMelt$variable, sep
= "_")

So I get a column in the appropriate format. But when I cast :

               go <- cast(genMelt,sample_id~Locus)

I get just the frequencies of the Loci per sample_id


Maybe I don't even need reshape. If you have any idea on this, I appreciate
it

Cheers,
Martin

	[[alternative HTML version deleted]]