search

Magdoll / Cogent

Coding Genome Reconstruction using Iso-Seq data
BSD 3-Clause Clear License
60 stars 17 forks source link

Raise CycleDetectedException #89

Open shri1984 opened 4 years ago

shri1984 commented 4 years ago

Hi, I am getting a errror "scipy.sparse.linalg.eigen.arpack.arpack.ArpackNoConvergence: ARPACK error -1: No convergence (77651 iterations, 66/100 eigenvectors converged)" when I rerun a one cluster after it failed in batch processing of bin.

I also got an error "CYCLE detected through simple_cycles! Raise CycleDetectedException!" when I was running a failed reconstruction (i used genome there). I followed the recommendations to fix failed runs. tried different k-mers. After a certain increase, I started to get the above error. I am using cogent 8.0.0. any idea how to fix this problem.

Magdoll commented 4 years ago

Hi @shri1984 , This probably means the input is just too repetitive - even with increasing K mer sizes, it's still blowing up the graph.

I can give it a shot if you are willing to share just that cluster but no guarantees it can be resolved.

-Liz

shri1984 commented 4 years ago

Hi,

Thanks for your reply. You are right. It is from an organism with the highly repetitive genome, I tried a few times now. It did not budge. I just wonder if I don't manage to fix those clusters is there any other way out to move forward?

Magdoll commented 4 years ago

Hi @shri1984 , Depends on what the goal is - What is the goal of using Cogent here? is it to do gene detection/annotation or create a mock reference genome?

shri1984 commented 4 years ago

Thanks again, We are building a transcriptome database covering as many genes as possible. I have a genome with lot of contigs (chromosome level as well as unplaced). Annotation is bad. So the plan is to use "annotated transcripts" to re annotate the genome and use that for other RNA seq projects. I found quite a few unmapped transcripts when mapped my HQ transcripts to that genome. That was surprising. So I started to look into those unmapped transcriptomes. That's how I started to use Cogent.
Also, another objective was to identify unique transcriptome models (denovo) and then get functional annotation and build pseudo genome to use as a reference (testing) and compare to the existing annotation.

Magdoll commented 4 years ago

Hi @shri1984 , Thanks for the response. That helps. Is it just this one cluster that had an issue? How many transcripts are in it?

Some idea for what to do if reconstruction fails - just use the transcripts as is. The downside is if you have a lot of transcripts in that partition/cluster it blows up the database size a bit.

If you are willing to share just this partition confidentially i can give it a try too.

-Liz