I have a dataframe with a first column contains the gene symbol and the others column contains an expression values. the Column of symbol can contain the same symbol more then one time.
So I would like for each set of line with the same symbol calculate the average (or median) of the lines. I will have only a single line by gene in the end.
Thank you in advance
library(dplyr)
dataframe %>%
group_by(Symbol) %>%
summarise(avg = median(expr.value))
I will use datatable if you have large matrix and you need speed. http://cran.r-project.org/web/packages/data.table/vignettes/datatable-intro.pdf .
exprmat_dt
Two other things to consider:
1) make sure you average log signal intensities and
2 ) make sure that similar gene symbols really refer to the same gene. For instance, if the data is from microarrays, the repplicate rows for the same gene symbol may be from different probes against the same gene but possibly obtained from different sequence databases or they are for different transcript variants. It might not be appropriate or sensible to average these values.
Aggregate is a nice solution, or you can use the very similar tapply:
res = tapply(yourDF$expression, yourDF$genesymbol, mean, na.rm=T)
They're practically identical in this case; the differences appear if you give them list of vectors to split on instead of just one. Aggregate will produce a data.frame with one column per factor plus one for the results; one row per combination, while tapply will produce an N-dimensional output (so columns are the first factor, rows the second ... and things get messier when N >= 3).
The output from aggregate is easier to feed into a number of functions (e.g. ggplot), while the tapply format might be a tad better if you'll be writing loops touching all possible combinations later. Aggregate also drops combinations with no data from the output, while tapply fills in with NAs. I also think the tapply table is nicer to read and present when there happens to be exactly 2 factors.
Thank you all for your responses!
sorry for the late reply
I was sick
Mohamed
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Statistical Software
- R