Compare Genomes

[infrared1983]

Hi everyone, I did Genome assembly by SOAPdenovo and obtained contigs and scaffolds, now I wanna compare these scaffolds to some reference genomes downloaded from Genbank. I am wondering what tools I can use to do this job, I used BLAST but it is not in good visualization, any other advice on such tools and programs?
I am new to this, any suggestion is appreciated.
Thanks

[jmw86069]

Use NUCmer from the MUMmer package (available from Sourceforge.) For large reference genomes, you have to split the reference into smaller pieces for it to fit in memory, then align everything to those pieces. It is relatively fast and efficient, and seems to be a commonly used tool as the basis for other assembly comparison pipelines. There are some recently available wrapper scripts which automate some of the steps, but it's not tough to run through the steps yourself. The online docs for MUMmer as a whole are very good.

[lh3]

For mammalian genomes, use bwa-sw. Two human assemblies in several CPU days.

[HerefordGuy]

De novo to reference comparison

I have had good luck with both MUMmer and MAUVE.
MAUVE is very straight forward and MUMmer I simply follow the recipes in the manuel.

[narain]

But is it good to use the contigs files or to use the scaffolds files for comparison?

[HerefordGuy]

Compare Genomes

I suppose it depends what your question is.
But, typically the goal of genome assembly is to create one scaffold per chromosome (assemble the entire genome/chromosome). With this in mind, I think it would make more sense to compare scaffolds to scaffolds.

[narain]

However, scaffolds are nothing but a combination of contigs filled with "estimates" of gaps in between. Does that not mean then that it is more sensible to compare contigs than to compare scaffolds as it has estimate of gaps in between which may be highly erroneous!