Difference between revisions of "Estimating Gene Count Talk"

From wiki
Jump to: navigation, search
 
Line 57: Line 57:
 
[[File:htcats.png]]
 
[[File:htcats.png]]
  
= Probabilistic approach =
+
= Probabilistic approach 1 =
  
Cufflink
+
* Cufflinks
 +
:- Reconstruct the transcripts from the data and annotation
  
cuffdiff
+
[[File:minpath.png]]
  
= Probabilistic approach =
+
= Probabilistic approach 2 =
 +
* Cuffdiff:
 +
:- Assign each read/fragment to a transcript with a probability maximum likelihood.
  
Cufflinks:
+
[[File:isolik.png]]
Reconstruct the transcripts from the data and annotation
 
 
 
= Probabilistic approach =
 
Cufflinks:
 
Reconstruct the transcripts from the data and annotation
 
 
 
Cuffdiff:
 
Assign each read/fragment to a transcript with a probability maximum likelihood.
 
 
 
= Probabilistic approach =
 
Cufflinks:
 
Reconstruct the transcripts from the data and annotation
 
Pros:
 
- Better methodology
 
- Integrated package (ease of use)
 
Cons:
 
Cuffdiff:
 
- Do not support alternative experiment design
 
- History of heterogeneous results/versions
 
* Assign each read/fragment to a transcript with a probability maximum likelihood.
 

Latest revision as of 13:34, 9 May 2017

Estimating Gene Count

How many reads are overlapping genomic features? - or - Can we confidently assign each read to a feature/transcript/gene? Not so simple.

We also have:

  • Multi mapping reads
  • Overlapping genes/transcripts

Two approaches:

  • Focus on what’s known with certainty
  • Probabilistic

Multi mapping reads

  • Unsolved problem:
- this can account for 10-30% of reads

Unsolved.png

  • Ignore them, but then again this decreases sensitivity
  • Weighted assignment

Of course, longer reads would solve this problem.

One transcript, one set of reads

T1.png

Two transcripts, another set of reads

T1t2.png

Aggregation to Gene-level 1

T1t2aggreg.png

Third transcript, another set of reads

T1t2t3.png

Aggregation to Gene-level 2

T1t2t3aggreg.png

HTSeq-count

  • Designed for RNA-Seq counting
  • Work at gene level
  • Remove multi-mapped reads
  • Several modes to resolve remaining uncertainty

Htseq.png

HTSeq-count modes

Htcats.png

Probabilistic approach 1

  • Cufflinks
- Reconstruct the transcripts from the data and annotation

Minpath.png

Probabilistic approach 2

  • Cuffdiff:
- Assign each read/fragment to a transcript with a probability maximum likelihood.

Isolik.png