定义选择数据的功能

Question

Let's start with my data.

    > dput(head(tbl_ready)) ## To make it clear I didn't put all of the row names
structure(list(Gene_name = structure(1:6, .Label = c("AT1G01050", 
"AT1G01080", "AT1G01090", "AT1G01220", "AT1G01320", "AT1G01420", 
"AT1G01470", "AT1G01800", "AT1G01910", "AT1G01920", "AT1G01960", 
"AT5G66570", "AT5G66720", "AT5G66760", "AT5G67150", "AT5G67360", 
"ATCG00120", "ATCG00160", "ATCG00170", "ATCG00190", "ATCG00380", 
"ATCG00470", "ATCG00480", "ATCG00490", "ATCG00500", "ATCG00650", 
"ATCG00660", "ATCG00670", "ATCG00750", "ATCG00770", "ATCG00780", 
"ATCG00800", "ATCG00810", "ATCG00820", "ATCG01090", "ATCG01110", 
"ATCG01120", "ATCG01240", "ATCG01300", "ATCG01310", "ATMG01190"
), class = "factor"), `10` = c(0, 0, 0, 0, 0, 0), `20` = c(0, 
0, 0, 0, 0, 0), `52.5` = c(0, 1, 0, 0, 0, 0), `81` = c(0, 0.660693687777888, 
0, 0, 0, 0), `110` = c(0, 0.521435654491704, 0, 0, 0, 1), `140.5` = c(0, 
0.437291194705566, 0, 0, 0, 1), `189` = c(0, 0.52204783488213, 
0, 0, 0, 0), `222.5` = c(0, 0.524298383907171, 0, 0, 0, 0), `278` = c(1, 
0.376865096972469, 0, 1, 0, 0), `340` = c(0, 0, 0, 0, 0, 0), 
    `397` = c(0, 0, 0, 0, 0, 0), `453.5` = c(0, 0, 0, 0, 0, 0
    ), `529` = c(0, 0, 0, 0, 0, 0), `580` = c(0, 0, 0, 0, 0, 
    0), `630.5` = c(0, 0, 0, 0, 0, 0), `683.5` = c(0, 0, 0, 0, 
    0, 0), `735.5` = c(0, 0, 0, 0, 0, 0), `784` = c(0, 0, 0.476101907006443, 
    0, 0, 0), `832` = c(0, 0, 1, 0, 0, 0), `882.5` = c(0, 0, 
    0, 0, 0, 0), `926.5` = c(0, 0, 0, 0, 1, 0), `973` = c(0, 
    0, 0, 0, 0, 0), `1108` = c(0, 0, 0, 0, 0, 0), `1200` = c(0, 
    0, 0, 0, 0, 0)), .Names = c("Gene_name", "10", "20", "52.5", 
"81", "110", "140.5", "189", "222.5", "278", "340", "397", "453.5", 
"529", "580", "630.5", "683.5", "735.5", "784", "832", "882.5", 
"926.5", "973", "1108", "1200"), row.names = c(NA, 6L), class = "data.frame")

Take a look on the names of the columns (just picked the 6 of them):

10
20
52.5
81
110
140.5

Those names tell me the size range. The size of the genes in the first column starts from 10 and ends on the begining of the second column = 20. That means that to the first column should belong genes with the size between 10-20.

I have another table which tells me what's the size of all genes (there are much more than can be finded in my first table):

    >dput(head(tbl_size))
    structure(list(Gene_name = structure(1:6, .Label = c("ATMG01290", "ATMG01300", "ATMG01310", "ATMG01320", "ATMG01330", 
    "ATMG01350", "ATMG01360", "ATMG01370", "ATMG01400", "ATMG01410"
    ), class = "factor"), tp = c(26L, 17L, 22L, 142L, 12L, 45L), 
        size = c(49.4255, 28.0913, 40.2872, 213.572, 24.4838, 70.4375
        )), .Names = c("locus", "tp", "size"), row.names = c(NA, 

6L), class = "data.frame")

and now the main part. What I want to achieve with my code ?

So, I'm trying to find only those genes which are found in the fractions (columns) with the size range two times higher than a real size of the gene. No idea if you understand what I am trying to do so let me use an example.

so let's say that we have a genes:

  Names      Size      
    AT1G01080     40
    AT1G01090     30
    AT1G01220     50

Let's multiply the size by 2:

    Names        Size      
    AT1G01080     80
    AT1G01090     60
    AT1G01220     100

In first table ( tbl_ready ) we can find the list of the genes and specific fractions (columns) defined by size which I explained in the begining of this thread. I would like to put the 0 instead of any values if any gene can be found in the fraction (column) which is not atleast two times higher than the gene size.

To find the size of the gene you have to look in the second table ( tbl_size ).

Just to sum it up. I'm trying to define which of those genes come atleast as a complex of 2. So only fractions with size two times higher than the size of the gene are important for me.

IF SOMEONE KNOWS WHAT I AM TRYING TO DO PLEASE EDIT MY QUESTION TO MAKE IT READABLE. I FEEL LIKE MY BRAIN IS DEAD.

Answer 1

Firstly, convert the columns to their numerical value:

frac <- as.numeric(colnames(tbl_ready))

and then get the index per gene of the column that doesn't exceed it's frac by two-fold:

ind <- lapply(tbl_size$size, function(x) which(frac > x*2)[1]-1)

Then you can create an array index of the values that you need to set to zero:

rowI = rep(match(tbl_size$locus, tbl_ready$Gene_name), times=ind-1)
colI = unlist(mapply(seq, from=2, length=ind-1))
tbl_ready[cbind(rowI, colI)] <- 0

You'll have to be careful if gene_names don't have a 1:1 mapping with locus, and cases where none of the columns exceed the gene size two fold, as there'll be NAs that need dealing with. I'm assuming you're stuck using these representations of your data, as it would probably be better to store tbl_ready in a longer narrower form than you have it here (containing only three columns name, size, and value - and omitted the zero values).

Answer 2

I'm going to change my original answer, this time using the data you've provided - the only real differences are that you've changed the column names (I'm assuming column tp in tbl_size is the thing we need to match to the column headings in tbl_ready), and that some of the rows in table_size don't correspond to tbl_ready.

Firstly, convert the columns to their numerical value:

frac <- as.numeric(colnames(tbl_ready))

and then get the index per gene of the column that doesn't exceed it's frac by two-fold:

mapToReady <- tbl_size$locus %in% tbl_ready[[1]]
ind <- sapply(tbl_size$tp[mapToReady], function(x) which(frac > x*2)[1]-1)

Then you can create an array index of the values that you need to set to zero:

rowI = rep(match(tbl_size$locus[mapToReady], tbl_ready[[1]]), times=ind-1)
colI = unlist(mapply(seq, from=2, length=ind-1))
tbl_ready[cbind(rowI, colI)] <- 0

So, for instance, AT1G01050 is the 5th row of tbl_size (none of the previous entries have an entry in your tbl_size), and the first row of tbl_ready. So the first 'iteration' of the sapply line hits 'tbl_size$tp[mapToReady][1]' which is the tp of AT1G01050 which is 12. 2*12 is 24, so is between 20.0 and 52.5, so we're going to need to set columns corresponding to '10', and '20' to zero, but not columns '52.5' onwards, for the AT1G01050. This corresponds to columns 2 and 3 of row 1 of tbl_ready, which is what the cbind portion of the last three lines is doing.