This project is designed to identify the mapping problem that occurs with STAR and our SUPT7L samples.
Many reads are not soft clipped instead of being split. Why is this the case? Is there a way to avoid this problem?
Region of interest: SUPT7L Observation: Soft-clipping at exon 3 instead of splitting the reads
TODO: Capture all values in a table to make it comparable
- Extract reads from region of interest that are soft clipped A. Reads aligned to SUPT7L ('2', 27873676, 27886481) B. Extract all reads from Exon3-Exon2 ('2', 27883851, 27885150) C. 4-base (CTGG) insertion in Exon3 D. Check if insertion is duplication E. Soft-clipping at Exon3 ('2', 27883851, 27884253): E.1. TTTGCAGATT in clipped sequence? E.2. TTTGC in clipped sequence? E.3. TTT in clipped sequence? F. Soft-clipping at Exon2 ('2', 27885044, 27885150) F.1. AGACGGAACT in clipped sequence? F.2. AGACG in clipped sequence? F.3. AGA in clipped sequence?
- Realign with same settings to validate
- Extract reads of interest
- Get statistics for reads of interest
- Apply changes to algorithm
- Collect statistics and compare
The problem is with the outSAMstrandField intronMotif, which was used to make the output Cufflinks/Cuffdiff compatible.