RStudio + sleuth! Error in if (sum(filter_bool) == 0) { : missing value where TRUE/FALSE needed

Question

[R]

hello my friends! i have a problem with Sleuth. RStudio output that error: Error in if (sum(filter_bool) == 0) { : missing value where TRUE/FALSE needed

sample_id <- dir(
                  file.path('~/Desktop/xxxxxx/kallisto_output')
)

kal_dirs <- file.path(
                  "~/Desktop/xxxxxx/kallisto_output", 
                  sample_id
)

s2c <- read.table(
    file.path("~/Desktop/xxxxxx/metadata/meta.tsv"), 
    header = TRUE, 
    stringsAsFactors = FALSE, 
    sep = " "
)

s2c <- dplyr::mutate(
                  s2c, 
                  path = kal_dirs
)

s2c$genotype_variation_s <- as.factor(
                  s2c$gene
)

s2c$genotype_variation_s <- relevel(
                  s2c$gene, 
                  ref = "c-like"
)

so <- sleuth_prep(s2c, full_model = ~gene,
                  read_bootstrap_tpm = TRUE, 
                  extra_bootstrap_summary = TRUE, 
                  num_cores = 4)


OUTPUT:
reading in kallisto results
dropping unused factor levels
...........................
normalizing est_counts
Error in if (sum(filter_bool) == 0) { : 
  missing value where TRUE/FALSE needed

---

s2c contain:
sample gene path
M_u_Kidney_clike c-like ~/path/to/kallisto/output
M_u_Kidney_x1 isoform ~/path/to/kallisto/output
M_u_Kidney_x2 isoform ~/path/to/kallisto/output
M_u_Kidney_x3 isoform ~/path/to/kallisto/output
M_u_Kidney_x4 isoform ~/path/to/kallisto/output
R_b_Kidney_clike c-like ~/path/to/kallisto/output
R_b_Kidney_x1 isoform ~/path/to/kallisto/output
R_b_Kidney_x2 isoform ~/path/to/kallisto/output
.................
R_b_Kidney_xX isoform ~/path/to/kallisto/output
_______________________________________________________________________________
27 objects

[dsull]

You probably aren't receiving any replies because your code is poorly formatted (it's very difficult to read it) and you aren't providing enough information (e.g. print out the first few lines of what each of your variables contain; what does t2g look like? what does s2c look like? etc.). Unless those things are fixed, it's hard for people to help you.

[R]

Hello my friend! thanks for answering! I redid the text to be more readable, I hope it got better.

[dsull]

There's only one sample in your "c-like" condition. That will be problematic for sleuth. You need replicates in all your conditions. You can't compare 1 sample with 26 other samples.

[R]

no, c-like 8 pieces.

for each tissue (kidney, skin, liver) 1 gene, somewhere even 2 genes

[dsull]

Thanks for clarifying. However, your entire path column has ~/path/to/kallisto/output for every single sample. Every sample should belong in a different directory.