Hi, I have one RSEM file. I am trying to import it with the following commands (R is invoked in the same directory of the input file):
library(tximport)
library(readr)
files <- c("Org_name_v1_Holo.RNA.fna.allSampless.RSEM_readCounts.results")
txi.rsem <- tximport(files, type = "rsem", txIn = FALSE, txOut = FALSE)
But I receive this ERROR Output:
reading in files with read_tsv
1
Error in computeRsemGeneLevel(files, importer, geneIdCol, abundanceCol, :
all(c(geneIdCol, abundanceCol, lengthCol) %in% names(raw)) is not TRUE
In addition: Warning message:
Unnamed col\_types should have the same length as col\_names. Using smaller of the two.
Potential issues
(after inspectig https://bioconductor.riken.jp/packages/3.7/bioc/vignettes/tximport/inst/doc/tximport.html#session-info):
* I am using one file only, is that the problem? Should one file be imported in a different way?
* If using only one file, should I us eit in a diff way?
* Maybe the file must be zipped?
* Maybe it must have this suffix: "genes.results.gz" ?
*Must one round the RSEM numbers before import with tximport?
Input file RSEM file header:
Name treat1Repaet1 treat1Repaet2 treat1Repaet3 treat2Repaet1 treat2Repaet2 treat2Repaet3 trea3Repaet1 treat3Repaet2 treat3Repaet3
Org1_PredGene_000123456.p1 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Org1_PredGene_000123457 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00
Org1_PASA_asmbl_000123458.p1 0.00 0.00 1.00 0.00 0.00 0.00 0.00 0.00 0.00 1.00
Org1_TRINITY_000123459 0.00 8.67 7.80 14.85 42.49 0.00 8.50 0.00 2.41 10.57
I am not sure I understood. Do you mean with single file I better not use tximport?
Instead I just read the file to an object with this R command: read.delim(file, header = TRUE, sep = "\t"....)
Yes. Tximport is really designed for aggregating data across samples and for summing to gene level. With only one sample, and using RSEM which outputs the gene summary file, there is no point to tximport.
I have many samples in the single input file. See matrix above, each column is a sample, each row a transcript id. Infect, I have much more columns than illustrated above. Actually the matrix contains: over 90 samples (columns) for over million transcripts (rows). However, I was suggested to use tximport since I have RSEM data. Would you recommend to use tximport for this case? Currently we are using isoform level counts.
You can just read in the data from the file. Someone must have compiled it. So don’t use tximport.
Dear Michael, thank you very much.
A colleague of mine generated the RSEM file. If there is any advantage for using tximport in our study, they can be changed to any other format.
However, it is important to note that we are using meta-transcriptome data (containing hosts, and microbiome etc'), and therefore we were concerned about unifying orthologs of several organisms into one gene, if we use gene level data.
Therefore the data produced by RSEM is isoform level counts .
In this regards, I have tried also to use the following command but received errors:
txi.rsem <- tximport(files, type = "rsem", txIn = FALSE, txOut = FALSE)
I'll just say, you can only use tximport to aggregate across many files, each with data for a single sample.
You have already aggregated RSEM output, which bypasses the need for tximport, so you don't need to use tximport.
If you want gene-level results use your own custom scripts on the gene-level output files from RSEM.
Thank you very much
Arik.