Trouble manipulating dataset

Question

[2405592M]

Hi Guys, I'm having trouble with some data manipulation. I have the following table where I have the raw counts from an RNA-seq experiment. I'm trying to group all my counts by tRNA wobble position.

head(df)

Isodecoder Anticodon Wobble Loci Fragment_type AAV_Ctrl1 AAV_Ctrl2 AAV_Ctrl3 AAV_Cre1 AAV_Cre2 AAV_Cre3
Ala       AGC      A    1   wholecounts         8         3         0       35       58       61
Ala       AGC      A   10   wholecounts         0         0         0        0        2        0
Ala       AGC      A   12   wholecounts         0         0         0        0        0        0
Ala       AGC      A    2   wholecounts      1228       839       766     1115     1525      784
Ala       AGC      A    3   wholecounts       286       125       120      504      387      380
Ala       AGC      A    4   wholecounts       675       541       353      328      452      367

I'm essentially trying to sum counts by Isodecoder and wobble position (i.e for Ala, I'd have 1 value for wobble A, T, C and G for each of my conditions Ctrl1-3 and Cre1-3). Tried to use dplyr with no luck. THANKS IN ADVANCE!!!

[russhh]

Could you show your dplyr attempt, please

Answer 1

[shawn.w.foley]

You can use aggregate through base R to accomplish this if you're just trying to sum by Isodecoder.

> df.agg <- aggregate(df[,-seq(5)],list(df$Isodecoder),sum)
> df.agg
  Group.1 AAV_Ctrl1 AAV_Ctrl2 AAV_Ctrl3 AAV_Cre1 AAV_Cre2 AAV_Cre3
1     Ala      2197      1508      1239     1982     2424     1592

The above command is making a new data frame called df.agg by taking all of the counts in df (removing the first five columns), aggregating it based on the values in df$Isodecoder, and performing the sum function.

[2405592M]

Perfect! Aggregate was the function I needed! Cheers!