Newbler sensitivity tuning for isoform detection?

[yvan.wenger]

Dear All,

I*performed an hybrid assembly with 454 Reads and Illumina-generated contigs (oases). I used the following command (mi 90, ml40 generated roughly the same output).

runAssembly -o hybrid_run_ml35_mi95 -ml 35 -mi 95 -cpu 0 -vt ./454adapters.fa -cdna NG-5322_454_1_sequence.sff NG-5322_454_2_sequence.sff ./v33_39_45.fa

I was wondering if parameters can be tuned so that the specific variations illustrated in the attached images can taken into account to produce more isotigs in this isogroup? I could not find any parameters (I simply tried then to go from ml40 mi90 to ml35 mi95 as it seems that variants are considered as sequencing errors, this did not work).

Ohterwise, intrestingly, the "hybrid" assembly (Illumina+454) produced only one isotig, whereas newbler 2.3 and 2.5 with 454 only produced 2 isotigs. I*am not quite sure why.

Thanks a lot for your insights!

Yvan

[flxlex]

I was wondering if parameters can be tuned so that the specific variations illustrated in the attached images can taken into account to produce more isotigs in this isogroup?

I am not sure I understand what you mean with producing more isotigs?

[yvan.wenger]

Hello flxlex,

I will try to be more clear. First, the two pictures I attached come from two isotigs (out of a total of two), corresponding to a single gene. There are no other isotigs/isogroups corresponding to this gene in the 454Isotigs.fna file. Finally, I could have attached only one of the pictures with the same question: both of the pictures illustrate an independent event where I think variations in reads exist, and are not otherwise reported in any consensus/isotig.

The two regions have a (albeit weak) support for alternative splicing, with 12bp missing (relative to the consensus) in some reads as shown in picture1 and an extra 53bp in picture 2 (here the read count is 3 with the stretch, 3 without the stretch). I know that these are variants actually existing in the transcripts, and I was wondering whether it was possible to increase newbler sensitivity so that these variants appear somewhere in the consensus generated, presumably as new isotigs in the 454Isotigs.fna file.

Thanks a lot for your help,

Yvan