Mann-Whitney U-test (utest)

This is a method for comparing datasets from a TnSeq library evaluated in two different conditions (i.e. pairwise-comparison, analogous to resampling). This is a non-parametric rank-based test on whether the level of insertions in a gene or chromosomal region are significantly higher or lower in one condition than the other. Effectively, the insertion counts at the TA sites in the region are pooled and sorted. Then the combined ranks of the counts in region A are compared to those in region B, and p-value is calculated that reflects whether there is a significant difference in the ranks.

The results (genes with adjusted P-value<0.05) should be similar to what you get with resampling. The advantage of this method is that it is less sensitive to outliers (e.g an unusually high insertion count at just a single TA site). A reference for this method is (Santa Maria et al., 2014).

Usage

> python3 transit.py utest <combined_wig_file> <metadata_file> <annotation_file> <condition_for_control> <condition_for_experimental> <output_file> [Optional Arguments]

Optional Arguments:
  --n <string>     :=  Normalization method. Default: --n TTR
  --iN <float>     :=  Ignore TAs occurring at given fraction (as integer) of the N terminus. Default: --iN 0
  --iC <float>     :=  Ignore TAs occurring at given fraction (as integer) of the C terminus. Default: --iC 0

GUI Mode

Utest can be access though the “Method” tab in the Menu Bar.

The parameters to input through the parameter panel for the method is equivalent to the command line usage, except in the GUI format we name the output files using the prefix passed in.

Output

The output file is tab-separated text file (spreadsheet) with the following columns:

Column Header	Column Definition
Orf	Gene ID.
Name	Name of the gene.
Desc	Gene description.
Sites	Number of TA sites in the gene.
Mean Ctrl	Mean of read counts in condition 1. (avg over TA sites and reps)
Mean Exp	Mean of read counts in condition 2.
Log 2 FC	Log-fold-change of exp (treatment) over ctrl (untreated)
U Statistic	test statistic reflecting which condition has higher counts
P Value	2-tailed P-value of u_stat based on Mann-Whitney
Adj P Value	Adjusted p-value controlling for the FDR (Benjamini-Hochberg)

The U-test statistic is a value reflecting whether insertion counts at TA sites in control samples are higher on average than counts in the experimental samples.

Important: The significant genes (conditionally essential) are those with Padj<0.05.

The significant genes should be comparable with those from resampling.

Runtime

The utest method is relatively fast, and should take less than a minute on most datasets.

---