Hello,
I am facing some wierd stuff with SPAdes v4.2.0 genome assemblies of fungal datasets.
Background.
From FastQ files after general QC using FastP i performed genome assemblies uisng SPAdes with following command (keeping same flags for all samples)
spades.py -k 21 33 55 77 99 111 127 --isolate -1 $read1 -2 $read2 -t $threads -o $asmOUT
I selected scaffolds.fasta for downstream analysis. removed scf <1000bp and ran genomes through NCBI-FCS to get clean assemblies.
Mapped QC-clean reads to genome and i get following alignment stats.
Weirdness i See
When i map the this genome assembly to reference using Nucmer and visualize the alignments as Circos i see chimeric scaffolds which show alignment to multiple Ref-Chromosomes. First i thought, it might be because i am using scaffolds.fasta from SPAdes, so i checked contigs.fasta and it also shows similar contigs. (this ruled out file selection issue). (Highlighted scaffolds)
2nd thing i think was Miss-assembly or something, so i ran Qualimap bamQC on bam file to see coverage across reference.
I also chekced individual Scf coverage and to me it looks OK (but i can be wrong in interpretation, so correct me kindly)
Now I have same genome also sequenced with PacBio HiFi and that assembly looks great with nothing like this (no chimeric chromosome stuff).
I have total 110 genome assemblies from Illumina data and ~63 of those show chimeric contigs like this. What can be issue here ? from coverage i understand that its normal, because these genomes are known to be repeat rich. but concerning point is that with pacbio, same genome doesnot show this phenomenon.
Your Input on this is highly helpfull.
Thank you.
0 answers
No answers yet.
Log in to answer this question.
-
Genome comperison showing unaligned sequence of 70kb in synteny
written by SomeOne •
-
Understanding IGV
written by zizigolu
-
How to do preprocessing from fastqc
written by Lulu •
-
Best practices in Fungal Genome Assembly
written by SomeOne •
-
nanopore data analysis methylation modification
written by Domi •
-
How to get sample details from NCBI datasets
written by beginner123 •
-
Why my `Percent duplicates` in a shallow sequencing is too high?
written by 631079064 •
-
Is the current R version (4.1.2) are not suitable for the packages related with TCGA
written by nonaddldy •
-
Can not plot degree distribution in Cytoscape Network Analyzer on logscale
written by bhodai •
-
Coverage plot/histogram of a complete genome
written by Michal Nevo •
My first guess would be repeat-resolving( or rather lack thereof with short reads)
with the long read approach it might be that some repeats are getting better resolved compared to using the short reads. With the short-read approach it is very likely that a "contig" ends in a repeat sequence (TE) and might thus get erroneously stitched to a next contig (that starts with a matching repeat) resulting in chimeric contigs/scaffolds.
Hi @lievensterck. Thank you for responce,
in this case, what would you suggest to resolve this issue, or just let them be and move on for downstream analysis ? These are Fusarium genomes.
If that is the case, this is a limitation of short read data. You could submit the illumina data as is to SRA (since it is already generated), making a note that using this data alone (with SPAdes and MaSUrCA) generates chimeric chromosome assemblies (indicate which chromosomes are affected).
what do you want to achieve in the end with these genomes? eg. if you are only interested in the gene-level analysis this will not hinder too much, if you for instance want to go into the direction of genome collinearlity then this will be a sever hurdle.
depending on the number of such cases I assume a manual intervention is perhaps not suitable?
Did you consider the hyrbrid assembly approach? for instance: using the long read to build a 'backbone' and use the illumina for polishing the base accuracy . There are nowadays several software that do this in sort of one procedure .
my initial plan is to go for gene level analysis along with marking COre and Accessory chromosomes. These choimeric contigs are coming up in Core chromosomes only. Unfortunately i dont have long read data for all 110 samples. i have it for just 10 samples, and i am used those for scaffolding of these illumina assemblies. (identifying closly related spp based on phylogenetic tree)