question about fragment size distribiton and error(NestedPathAbundanceEstimator) when using short read illumina RNA seq data.

[jjuhyunkim]

Hi,

I have questions about fragment size using short read data. When I analyzed short read paired Illumina RNA data, the RPVG estimated the fragment distribution with a mean of 181.201 and a standard deviation of 54.2619. Should this estimation represent the whole paired-end fragment size, which corresponds to the cDNA library length, rather than the single-end read length? If so, I would expect the fragment mean size to be around 300-500, and I plan to use a mean of 500.

Additionally, I encountered an error message: NestedPathAbundanceEstimator::inferPathSubsetAbundance const: Assertion path_group.second.size() <= group_size' failed.` Could this error be related to the issue I mentioned above?

Thank you!

[zwh82]

Additionally, I encountered an error message: NestedPathAbundanceEstimator::inferPathSubsetAbundance const: Assertion path_group.second.size() <= group_size' failed.`

I have encountered the same problem, have you solved it? My read distribution estimate is correct, it should be unrelated to the problem.

I checked all path_group. This situation only occurs twice, which should not be a common occurrence. I'm not sure what exactly caused it.

Path group ID: 2092, Size: 3, Group size limit: 2, Paths: 0 1 2 
Skipping path group due to size limit.
Path group ID: 2093, Size: 1, Group size limit: 2, Paths: 3 
Path group ID: 2094, Size: 3, Group size limit: 2, Paths: 4 5 6 
Skipping path group due to size limit.

[jjuhyunkim]

I couldn't resolve this yet, what was your read length estimation? Did you use illumina short read data?

[jjuhyunkim]

Hi @jeizenga,
Kind reminder regarding this issue😁 Could you please leave a comment if you have any ideas?

This is what error message states:

Running rpvg (commit: 301f553412a7f3b3c3dccad74e845868da4f0468)
Random number generator seed: 1730338817
Fragment length distribution parameters found in alignment (mean: 185.155, standard deviation: 61.6908)
Loaded graph, GBWT and r-index (6.56369 seconds, 2.55947 GB)
Fragment length distribution parameters re-estimated from alignment paths (location: 150.217, scale: 90.2194, shape: 78.2476)
Found alignment paths (3086.88 seconds, 2.55947 GB)
Clustered alignment paths (0.913452 seconds, 2.55947 GB)
rpvg: /home/rpvg/src/path_abundance_estimator.cpp:715: void NestedPathAbundanceEstimator::inferPathSubsetAbundance(PathClusterEstimates*, const std::vector<ReadPathProbabilities>&, std::mt19937*, const spp::sparse_hash_map<std::vector<unsigned int>, double>&) const: Assertion `path_group.second.size() <= group_size' failed.
/var/spool/slurm/slurmd/job39745670/slurm_script: line 4: 3779436 Aborted                 (core dumped) rpvg -t 50 --graph pantranscriptome.xg --paths pantranscriptome.gbwt --alignments aligned.gamp --output-prefix test -f pantranscriptome.txt --inference-model haplotype-transcripts

My alignmnet data is Illumnia short read paired RNA seq data and the graph was generated by cactus-pangenome using assemblies (2 haplotype from one sample and one haplotype of chm13 reference).