Hi,
I have been using Transrate 1.0.3 to benchmark my RNA-seq assemblies.
I noticed that a more contiguous assembly would have very low (eg. ~0.1) p fragments mapped in the log, but Transrate's BAM file has over 90% properly mapped read pairs according to samtools flagstat. A less contiguous assembly would have about 0.7 p fragments mapped. I have also tried a reference-based approach such as rnaQUAST and it seems to prefer the more contiguous assembly over a less contiguous one. I am really puzzled by this. Is Transrate not able to handle too many "long" isoforms (ie. when many read pairs are multi-mapped across multiple isoforms)?
Hi, I have been using Transrate 1.0.3 to benchmark my RNA-seq assemblies. I noticed that a more contiguous assembly would have very low (eg. ~0.1)
p fragments mappedin the log, but Transrate's BAM file has over 90% properly mapped read pairs according tosamtools flagstat. A less contiguous assembly would have about 0.7p fragments mapped. I have also tried a reference-based approach such as rnaQUAST and it seems to prefer the more contiguous assembly over a less contiguous one. I am really puzzled by this. Is Transrate not able to handle too many "long" isoforms (ie. when many read pairs are multi-mapped across multiple isoforms)?