Hello, I have been using txImport to add up all the transcripts to get one TPM value per gene. When I add up genes manually in excel, it doesn't always match tximport's TPM value. Why is tximport TPM's sometimes less than the expected value I calculated in Excel? Here is a comparison of the TPMs:

Gene- [Excel] vs [TxImport Result]

Aanat - 0 vs 0

Aatk - 0.29 vs 0.17

Abca1 - 3.28 vs 3.28

Abca4 - 0.25 vs. 0.22

Abca2 - 3.2 vs 2.75

Abcb7 - 0.95 vs 0.95

Manual Count in Excel:

TxImport Result:

Here is my code for reference

#load in data
#make file path to file in C:\Users\cathe\Desktop
files <- file.path("C:", "Users", "cathe", "Desktop", "GSM3106294_MEF_1_quant2.sf.txt", fsep="/")

#get gene ID's and TX names
library(TxDb.Mmusculus.UCSC.mm10.knownGene)
txdb <- TxDb.Mmusculus.UCSC.mm10.knownGene
txdb
k <- keys(txdb, keytype = "GENEID")
length(k)
df <- select(txdb, keys = k, keytype = "GENEID", columns = "TXNAME") 
tx2gene <- df[, 2:1]  # tx ID, then gene ID
head(tx2gene)

#run tximport
txi.salmon <- tximport(files, type = "salmon", tx2gene = tx2gene)
txi.salmon
#view tximport output that has gene level quantification
head(txi.salmon$counts)

#convert the tximport object to a data frame for downstream operations
txi.salmon <- as.data.frame(txi.salmon)
#move rownames with the gene ID to a column using data.table package
setDT(txi.salmon, keep.rownames = "entrezgene_id")
#convert entrezid from character to int data type downstream operations
txi.salmon$entrezgene_id <- as.integer(txi.salmon$entrezgene_id)

#get gene symbol for entrez id in another data frame 
library(biomaRt)
mart <- useMart('ensembl', dataset = 'mmusculus_gene_ensembl')
entrezid2symbol <- 
  getBM(
    attributes = c(
      'ensembl_gene_id',
      'entrezgene_id', 'mgi_symbol'),
    filters = 'entrezgene_id',
    values = txi.salmon$entrezgene_id, 
    mart = mart)
entrezid2symbol <- as.data.frame(entrezid2symbol)

#merge both tximport salmon output and entrezid2symbol to include gene symbols
salmonids <- merge(txi.salmon,entrezid2symbol,by=c("entrezgene_id"))

write.table(salmonids, file="output.txt", sep = "\t", row.names=FALSE,  col.names=FALSE, quote = FALSE)


sessionInfo( )
> sessionInfo()
R version 4.0.5 (2021-03-31)
Platform: x86_64-w64-mingw32/x64 (64-bit)
Running under: Windows 10 x64 (build 18363)

Matrix products: default

locale:
[1] LC_COLLATE=English_United States.1252  LC_CTYPE=English_United States.1252   
[3] LC_MONETARY=English_United States.1252 LC_NUMERIC=C                          
[5] LC_TIME=English_United States.1252    

attached base packages:
[1] parallel  stats4    stats     graphics  grDevices utils     datasets  methods   base     

other attached packages:
 [1] data.table_1.14.0                         TxDb.Mmusculus.UCSC.mm10.knownGene_3.10.0
 [3] GenomicFeatures_1.42.3                    GenomicRanges_1.42.0                     
 [5] GenomeInfoDb_1.26.7                       biomaRt_2.46.3                           
 [7] org.Mm.eg.db_3.12.0                       AnnotationDbi_1.52.0                     
 [9] IRanges_2.24.1                            S4Vectors_0.28.1                         
[11] Biobase_2.50.0                            BiocGenerics_0.36.0                      
[13] readr_1.4.0                               tximport_1.18.0                          
[15] MOFA2_1.3.0                              

loaded via a namespace (and not attached):
 [1] bitops_1.0-6                matrixStats_0.58.0          bit64_4.0.5                
 [4] filelock_1.0.2              RColorBrewer_1.1-2          progress_1.2.2             
 [7] httr_1.4.2                  tools_4.0.5                 utf8_1.2.1                 
[10] R6_2.5.0                    HDF5Array_1.18.1            uwot_0.1.10                
[13] DBI_1.1.1                   colorspace_2.0-0            rhdf5filters_1.2.0         
[16] withr_2.4.2                 tidyselect_1.1.0            prettyunits_1.1.1          
[19] bit_4.0.4                   curl_4.3                    compiler_4.0.5             
[22] basilisk.utils_1.2.2        xml2_1.3.2                  DelayedArray_0.16.3        
[25] rtracklayer_1.49.5          scales_1.1.1                askpass_1.1                
[28] rappdirs_0.3.3              Rsamtools_2.6.0             stringr_1.4.0              
[31] basilisk_1.2.1              XVector_0.30.0              pkgconfig_2.0.3            
[34] MatrixGenerics_1.2.1        dbplyr_2.1.1                fastmap_1.1.0              
[37] rlang_0.4.10                rstudioapi_0.13             RSQLite_2.2.5              
[40] generics_0.1.0              jsonlite_1.7.2              BiocParallel_1.24.1        
[43] dplyr_1.0.5                 RCurl_1.98-1.3              magrittr_2.0.1             
[46] GenomeInfoDbData_1.2.4      Matrix_1.3-2                Rcpp_1.0.6                 
[49] munsell_0.5.0               Rhdf5lib_1.12.1             fansi_0.4.2                
[52] reticulate_1.18             lifecycle_1.0.0             stringi_1.5.3              
[55] SummarizedExperiment_1.20.0 zlibbioc_1.36.0             rhdf5_2.34.0               
[58] Rtsne_0.15                  plyr_1.8.6                  BiocFileCache_1.14.0       
[61] grid_4.0.5                  blob_1.2.1                  ggrepel_0.9.1              
[64] forcats_0.5.1               crayon_1.4.1                lattice_0.20-41            
[67] Biostrings_2.58.0           cowplot_1.1.1               hms_1.0.0                  
[70] pillar_1.6.0                reshape2_1.4.4              XML_3.99-0.6               
[73] glue_1.4.2                  vctrs_0.3.7                 gtable_0.3.0               
[76] openssl_1.4.3               purrr_0.3.4                 tidyr_1.1.3                
[79] assertthat_0.2.1            cachem_1.0.4                ggplot2_3.3.3              
[82] tibble_3.1.0                pheatmap_1.0.12             GenomicAlignments_1.26.0   
[85] memoise_2.0.0               corrplot_0.84               ellipsis_0.3.1

Sorry I missed this, the email notifications are broken it seems.

Are you using a tx2gene table that matches your transcripts? It needs to be exactly the same (e.g. all isoforms are present in the table).

How did you quantify the abundance -- what reference transcripts were used?

Secondly -- if you are summarizing in Excel using a different isoform to gene mapping, you will get a different result when you use a different mapping for tximport.