Differential Expression

Detecting Differentially Expressed Sequences

To run differential expression users must first configure the offset values. After this users must select a fold change cutoff value. No sequences with a lower value than this (default 1.0) will be reported.

No sequences with a fold change cutoff lower than this will be reported
No sequences with a fold change cutoff lower than this will be reported

 

Users will then select which samples will be compared against each other. This will involve selecting one as a reference and the other as the observed. For example, a wild type sample will be the reference and a particular treatment will be the observed. Or in a time series, time point 1 will be compared against time point 2.

A list of all samples is given in the initial table, users must drag the required sample names into the reference and observed lists.

Differentially expressed sequences are displayed as a table containing the following values:

  • Sequence – the RNA sequence itself
  • Size – the size class
  • Annotation – the type of annotation (for example miRNA)
  • Fold Change – the calculated fold change for that sequence (given in log2)
  • Direction – the direction of expression (up or down regulated)
  • Reference Count – the original count of that sequence in the reference sample
  • Observed Count – the original count of that sequence in the observed sample
  • Average Count  – the average count (log of the average of counts from the two samples)
  • Pattern – the pattern the sequence forms across the entire experiment. For example, with 3 time point samples, if a sequence was up-regulated from the first to the second, then down-regulated in the third, the pattern would be {UD}
Each predicted differentially expressed sequence is reported
Each predicted differentially expressed sequence is reported

When result sets are extremely large users must move the blue slider to view windows of results broken into chunks of 10000 sequences. Largely for performance issues.

Users can search the current window of results for a certain sequence, if they find it and notice it has been annotated as something (for example a miRNA) they can then select the sequence with the check box on the first column and click the view annotations buttons to show all annotations for that sequence

A single sequence may align in many places and consequently have many potential annotations. This table shows all of them
A single sequence may align in many places and consequently have many potential annotations. This table shows all of them

A suite of tools for analysing micro RNA and other small RNA data from High-Throughput Sequencing devices