Genome Assembly Sorting

Question

[Umer]

HI,

I have generated some genome assemblies using Nanopore data with FLYE assembler.

I have also performed 3x rounds of polishing with Racon (using nanopore data) and 5x rounds with Pilon (using illumina data)

Now t have looked at my fasta files and they are not sorted (as longest to shortest contigs) and the contig numbering is also not in numerical order.

>contig_1
>contig_102
>contig_103
>contig_104
>contig_105
>contig_106
>contig_107
>contig_11
>contig_110
>contig_111
>contig_112
>contig_113
>contig_114
>contig_117
>contig_120
>contig_121
>contig_122
>contig_124
>contig_125
>contig_128

My Questions

1. Do I need to perform this contig renaming and sorting at the first step where i get the genome assemblt fasta files?
2. Or I should sort the final polished assembly (longest to shortest) and rename them ?
3. Is their any specific way to raname contigs ?

[Joe]

If you look, your contigs do appear to be sorted by numerical value, its just not a "natural" sort because the numbers are not to the same significant figures or zero-padded.

As genomax points out though, this rarely matters.

[Umer]

Hi, I understand your point, but if you see the above shared answern this is what i mean.

Answer 1

[GenoMax]

This is --> https://stackoverflow.com/questions/45950646/what-is-lexicographical-order

Why are you bothered by it? If you want to sort your contigs by length then use --> Order contig by size

[Umer]

HI, the reason I am concerned is that

When i generate samtools faidx index of original assembly they look like this

contig_1    21276   10  60  61
contig_102  5361    21653   60  61
contig_103  7061    27116   60  61
contig_104  121625  34307   60  61
contig_105  41647   157972  60  61
contig_106  59630   200326  60  61
contig_107  62762   260962  60  61
contig_11       216926  324782  60  61
contig_110  11781   545336  60  61
contig_111  1667    557326  60  61
contig_112  2011    559033  60  61

when i sort the assembly.fasta via seqkit sort -l -r assembly.fasta > assembly_sort.fasta and then index using samtools the contigs rearrange according to length like this

contig_43   6501589       11    60  61
contig_5    5413231 6609970         60  61
contig_18   5374465 12113433    60  61
contig_2    5140464 17577483    60  61
contig_4    4441101 22803632    60  61
contig_34   4232623 27318763    60  61
contig_28   3504136 31621941    60  61
contig_17   3273534 35184491    60  61
contig_47   3179198 38512595    60  61
contig_32   2708330 41744791    60  61
contig_25   2470542 44498271    60  61

so this made me think i i have to sort the assembled genome, if yes, then before polishing steps or after polishing its also OK ?

[GenoMax]

You are using seqkit with options to sort the data based on length (-l) and reverse sort (-r). So it is not surprising that you are getting a different list of contigs. If you used -n (or just did seqkit sort), your sort result should look identical.

This is simply rearranging the data according to its length without changing it.

Answer 2

[colindaven]

I'd use a tool like RagTag if you have a related genome to scaffold the contigs into pseudochromosomes.

https://github.com/malonge/RagTag

Then you don't need to worry too much about contig names.