Chloroplast google sheet
Files archived at: /redser4/projects/dittrichia/organelles/chloroplast
Chloroplast genome in: /redser4/projects/dittrichia/organelles/chloroplast/flye/ditt_lim_flye_chloroplast_adapterfilt_edited2.tar.gz
Initial attempts were made with GetOrganelle v1.7.4 using get_organelle_from_assembly.py and the assembly graph output from HiFiasm. Results were very fragmented.
Script: assemblies/getorganelle/getorg.sh
HiFi reads were then aligned to the chloroplast assembly from the closest available relative, Limbarda crithmoides. Unmapped reads (-F 4) and secondary aligments (-F 256) were filtered from the bam output.
Script: minimap.sh
Read ids remaining after filtering were used to subset the adapter-filtered HiFi reads.
Script: seqtk.sh
This subset of reads was used for chloroplast assembly, testing Canu v2.2, HiFiasm v0.16.1-r375, and Flye v2.9. Because HiFiasm and Flye output assembly graphs, those were also run through GetOrganelle. Flye/GetOrgnaelle produced an assembly that looked complete but with extra contigs of much lower coverage, possibly segments that also closely match regions within the nuclear genome, or maybe artifacts of other duplication. I manually removed those from the graph and re-ran GetOrganelle on that graph which successfully created a circular genome.
The Limbarda chloroplast assembly was mapped to the Dittrichia nuclear assembly with blastn. Three more substantial regions matched, indicating the potential for chloroplast mis-assembled in the nuclear assembly. HiFi reads used for assembly were aligned to the assembly to check for issues with mapping or coverage in IGV.
Scripts: ../alignments/minimap/self/minimap.sh, ../alignments/minimap/self/filterFasta.sh, ../alignments/minimap/self/createFasta.py
