I am new to R and I don't quite know what structures to use and the correct syntax for them.
I have lists (that are more like tables with columns and column names). I would like to do the same functions to multiple lists. I assumed for loops would be reasonable to use.
My functions are
1) use a column to calculate a new column. (calculate fold change from log2foldchange)
2) make a new list using a subset of the old list and name it adjusting the name of the original list name
Here are the lines of code that worked for these tables individually.
#take values from the log2FoldChange column and calculate Fold Change
resCondition_anno$FoldChange <- 2^resCondition_anno$log2FoldChange
#subset my dataset based on the values for each row in the padj column
resCondition_anno_padj05 <- subset(resCondition_anno, resCondition$padj <= 0.05)
I would like to do these functions to multiple tables.
When I tried to do it in a for loop
resfiles1 <- c(resCondition_anno,resVirus_anno,resInter_anno)
for (i in resfiles1){
i$FoldChange <- 2^i$log2FoldChange # I was trying to calculate a new column based on log2FoldChange column
i_with_padj05 <- paste(i,"_padj05") # I was trying to create a new name like resCondition_anno_padj05
i_with_padj05 <- subset(i, i[[padj]] <= 0.05) # I was trying to subset my dataset based on values in the padj column
}
I tried to access the columns of my tables with $ and that gave me
Error: $ operator is invalid for atomic vectors
I tried to access the columns of my tables with [padj], I get
Error in subset.default(i, i[padj] <= 0.05) : object 'padj' not found
When I tried to access the columns of my table with `[[padj]], I got the following error
Error in subset.default(i, i[[padj]] <= 0.05) : object 'padj' not found
Am I going about this completely the wrong way? Is for loops reasonable way to approach my goals? I know apply functions exists but I had such a hard to getting output files out of them when I tried to input multiple files into it so I wanted to give for loops a try.
I would appreciate a code that would work for a random table and does these things and then I can figure out whether my tables are weird.
dput(head(resCondition_anno))
structure(list(ensembl = c("ENSMUSG00000051951", "ENSMUSG00000102331",
"ENSMUSG00000025902", "ENSMUSG00000104238", "ENSMUSG00000102269",
"ENSMUSG00000096126"), baseMean = c(2.34691358937965, 0.169507902147731,
49.4591642836684, 0.253911076708937, 3.27439052075304, 0.258178295608587
), log2FoldChange = c(1.04699290132002, 1.89907052894015, 0.629095304499277,
0.0597400040882164, -0.291997327218544, 1.97984690635658), lfcSE = c(1.09309963258445,
4.36961772602319, 0.291712394209747, 4.37647193807779, 1.21524080418346,
4.3263845102792), stat = c(0.95782019324678, 0.434607933236415,
2.15656008104662, 0.0136502655411644, -0.240279396654017, 0.457621577937096
), pvalue = c(0.338153434807336, 0.66384703564954, 0.0310399577136823,
0.989109002094381, 0.810113666298446, 0.647224338296786), padj = c(NA,
NA, 0.106540309680362, NA, 0.911344697137259, NA), mgi_symbol = c("Xkr4",
"Gm19938", "Sox17", "Gm37587", "Gm7357", "Gm22307"), gene_biotype = c("protein_coding",
"sense_intronic", "protein_coding", "processed_transcript", "processed_pseudogene",
"snRNA")), class = c("data.table", "data.frame"), row.names = c(NA,
-6L), .internal.selfref = <pointer: 0x0000027bef7e1ef0>)`
Expected results for the aim 1
> dput(head(resCondition_anno))
structure(list(ensembl = c("ENSMUSG00000051951", "ENSMUSG00000102331",
"ENSMUSG00000025902", "ENSMUSG00000104238", "ENSMUSG00000102269",
"ENSMUSG00000096126"), baseMean = c(2.34691358937965, 0.169507902147731,
49.4591642836684, 0.253911076708937, 3.27439052075304, 0.258178295608587
), log2FoldChange = c(1.04699290132002, 1.89907052894015, 0.629095304499277,
0.0597400040882164, -0.291997327218544, 1.97984690635658), lfcSE = c(1.09309963258445,
4.36961772602319, 0.291712394209747, 4.37647193807779, 1.21524080418346,
4.3263845102792), stat = c(0.95782019324678, 0.434607933236415,
2.15656008104662, 0.0136502655411644, -0.240279396654017, 0.457621577937096
), pvalue = c(0.338153434807336, 0.66384703564954, 0.0310399577136823,
0.989109002094381, 0.810113666298446, 0.647224338296786), padj = c(NA,
NA, 0.106540309680362, NA, 0.911344697137259, NA), mgi_symbol = c("Xkr4",
"Gm19938", "Sox17", "Gm37587", "Gm7357", "Gm22307"), gene_biotype = c("protein_coding",
"sense_intronic", "protein_coding", "processed_transcript", "processed_pseudogene",
"snRNA"), FoldChange = c(2.0662186086592, 3.72972827627808, 1.54659483966075,
1.0422779093498, 0.816770504282921, 3.94451221821964)), class = c("data.table",
"data.frame"), row.names = c(NA, -6L), .internal.selfref = <pointer: 0x0000027bef7e1ef0>)
Expected results for aim2
> dput(head(resCondition_anno_padj05))
structure(list(ensembl = c("ENSMUSG00000103922", "ENSMUSG00000025907",
"ENSMUSG00000061024", "ENSMUSG00000025911", "ENSMUSG00000025935",
"ENSMUSG00000025937"), baseMean = c(7.45083924607695, 1035.42915800337,
756.089939474399, 1510.50670239711, 2014.55644970672, 5206.99654662079
), log2FoldChange = c(3.31157886392159, -0.345358245876914, 0.340037961752993,
-0.637902858828505, 0.592795289538968, 0.59912370697665), lfcSE = c(0.984296895396084,
0.131191642000487, 0.0967702378760271, 0.120687031774959, 0.114283891072725,
0.161639505766009), stat = c(3.36441055479404, -2.63247140298489,
3.51386923517349, -5.28559572181691, 5.18704153292907, 3.70654255676794
), pvalue = c(0.000767073434065771, 0.00847661586751943, 0.000441630160084079,
1.25296333033368e-07, 2.13661093734535e-07, 0.000210107944374613
), padj = c(0.00522376704325313, 0.0385092726153939, 0.00325683272694307,
2.17721401368104e-06, 3.51690667040699e-06, 0.00168321660710376
), mgi_symbol = c("Gm6123", "Rb1cc1", "Rrs1", "Adhfe1", "Tram1",
"Lactb2"), gene_biotype = c("processed_pseudogene", "protein_coding",
"protein_coding", "protein_coding", "protein_coding", "protein_coding"
), FoldChange = c(9.92852128160573, 0.787112498791522, 1.26578990036559,
0.642646438673565, 1.5081660610658, 1.51479619975327)), class = c("data.table",
"data.frame"), row.names = c(NA, -6L), .internal.selfref = <pointer: 0x0000027bef7e1ef0>)
for aim 1
library(dplyr)
resCondition_anno_dumb <- resCondition_anno # produce a similar list
resCondition_anno_dumb$log2FoldChange <- resCondition_anno$log2FoldChange*3 # make some changes
list_t <- list(resCondition_anno, resCondition_anno_dumb) # here you enter your dataframes
# mutate adds a column to existing data sets, lapply makes it recursive
new_list <- lapply(list_t, function(x){x %>% mutate(FoldChange=2^log2FoldChange)})
for aim 2 something like
new_list <- lapply(list_t, function(x){x %>% filter(padj<=0.05)})
or you can pipe them together:
new_list <- lapply(list_t, function(x){x %>% mutate(FoldChange=2^log2FoldChange) %>% filter (padj <=0.05)})