ImpulseDE2 - help!

Question

[jsena]

Hi,

I have attempted to run ImpulseDE2 (R package) for my Small RNA-Seq time series data. I have checked that my counts data matrix and sample information data frame contain the same names. However, I receive the following error.

Error in .validate_names(colnames, ans_colnames, "assay colnames()", "colData rownames()") : 
  assay colnames() must be NULL or identical to colData rownames()

Could this be due to the way that I am importing my read counts matrix or sample data frame? Any advice that you can give will be greatly appreciated!

My code is below:

library(ImpulseDE2)

# Read Data into R as a table
 countdata <- read.table("counts_sncRNA_DE.tsv", header=T, sep='\t',row.names=1)

# Convert table into a counts matrix
 countdata <- as.matrix(countdata)

# Import my time series metadata into R and convert to data frame
 annot <- read.table("coldata_3.txt", header=TRUE, sep='\t')

annot <- as.data.frame.matrix(annot)

# DEG analysis with ImpulseDE2
 impulse_results <- runImpulseDE2(matCountData = countdata,
                                     dfAnnotation = annot,
                                     boolCaseCtrl = TRUE,
                                     vecConfounders = c("Batch"),
                                     scaNProc = 2,
                                     scaQThres = 0.05,
                                     boolIdentifyTransients = TRUE)

Process input
Processing Details:
ImpulseDE2 runs in case-ctrl mode.
Found time points: 2,6,16
Case: Found the samples at time point 2: M1C_2hr,M2C_2hr,M3C_2hr,M4C_2hr
Case: Found the samples at time point 6: M1C_6hr,M2C_6hr,M3C_6hr,M4C_6hr
Case: Found the samples at time point 16: M1C_16hr,M2C_16hr,M3C_16hr,M4C_16hr
Control: Found the following samples at time point 2:M1D_2hr,M2D_2hr,M3D_2hr,M4D_2hr
Control: Found the following samples at time point 6:M1D_6hr,M2D_6hr,M3D_6hr,M4D_6hr
Control: Found the following samples at time point 16:M1D_16hr,M2D_16hr,M3D_16hr,M4D_16hr
Found the following samples for confounder Batch and batch M1: M1C_2hr,M1C_6hr,M1C_16hr,M1D_2hr,M1D_6hr,M1D_16hr
Found the following samples for confounder Batch and batch M2: M2C_2hr,M2C_6hr,M2C_16hr,M2D_2hr,M2D_6hr,M2D_16hr
Found the following samples for confounder Batch and batch M3: M3C_2hr,M3C_6hr,M3C_16hr,M3D_2hr,M3D_6hr,M3D_16hr
Found the following samples for confounder Batch and batch M4: M4C_2hr,M4C_6hr,M4C_16hr,M4D_2hr,M4D_6hr,M4D_16hr
Input contained 667 genes/regions.
Selected 667 genes/regions for analysis.

Run DESeq2: Using dispersion factorscomputed by DESeq2.
<simpleError in .validate_names(colnames, ans_colnames, "assay colnames()", "colData rownames()"): assay colnames() must be NULL or identical to colData rownames()>
[1] "WARNING: DESeq2 failed on full model - dispersions may be inaccurate.Estimating dispersions on reduced model. Supply externally generated dispersion parameters via vecDispersionsExternal if there is a more accurate model for your data set."
Error in .validate_names(colnames, ans_colnames, "assay colnames()", "colData rownames()") : 
  assay colnames() must be NULL or identical to colData rownames()
In addition: Warning message:
In value[[3L]](cond) :
  Warning generated in dispersion factor estimation. Eead stdout.
Timing stopped at: 0.072 0.004 0.077
Timing stopped at: 0.495 0.012 0.507

Answer 1

[jsena]

I corrected this error.

First, I added a column before the sample column in my data frame. Previously, my data frame only contained the Sample, Condition, Time and Batch columns. The data frame should resemble below

 head(annot, 4)
         Sample Condition Time Batch
M1C_2hr M1C_2hr      case    2    M1
M2C_2hr M2C_2hr      case    2    M2
M3C_2hr M3C_2hr      case    2    M3
M4C_2hr M4C_2hr      case    2    M4

Secondly, I imported by data frame like so.

annot <- read.table("coldata.txt", header=TRUE, sep='\t', row.names=1)
annot <- as.data.frame.matrix(annot)

[Mahdiar Sadeghi]

I have had the same experience, the row.names() should be the same as sample id, and more interestingly it is better to sort the time-points before making ImpulseDE object.