Typical assembly wokflow for ALLHiC scaffolding

[baozg]

Hi @tangerzhang ,

I have a high heterozygosity polypoidy genome to assembly. After reading the manusrcript of ALLHiC and issues in the github, I want to confirm the typical workflow for ALLHiC scaffoling assembly. Here is my thoughts. Do you think there are any step need to be added or corrected?

• Using Canu to assembly allele contig by PB / ONT reads (100x for the ploypoidy genome size) . The parameters of Canu should be the canu-poly or others ?
• purge_haplotigs to remove alternative haplotigs
• 10x , BioNano, BAC etc. to improve the contig assembly
• 3D-DNA pipeline to correct the misassembly in the contig
• repeat masking and gene annotation on the contig assembly
• get allelic table by blastn to the progenitor diploid genomes speceies (If have two progenitor diploid genomes , do I need align to the different genome to get two allleic table? )
• ALLHiC pipeline

Best,
Zhigui

[tangerzhang]

Hi Zhigui,
Please see the following message for my suggestion:
Using Canu to assembly allele contig by PB / ONT reads (100x for the ploypoidy genome size) .

- Yes, please use canu to assemble the allele contigs. Pacbio reads are highly recommended. I have not tested ONT reads and have no idea how ONT will perform on polyploid genome.

The parameters of Canu should be the canu-poly or others ?

- Using canu-poly parameters will get better results, but this step is time-consuming. I do not recommend canu-poly parameters if your genome size is big (let's say more than 2 Gb).

purge_haplotigs to remove alternative haplotigs

- ALLHiC is designed for haplotype phasing and our program will perform better if haplotigs are assembled as many as possible. Therefore, please DO NOT use purge_haplotigs to remove alternative haplotigs.

10x , BioNano, BAC etc. to improve the contig assembly

- 10x, BioNano and BAC will definitely improve contig continuity but may introduce assembly error. I do not have much experience on 10x, Bionano assembly.

3D-DNA pipeline to correct the misassembly in the contig

- Correction of misassembled contigs will improve Hi-C scaffolding.

repeat masking and gene annotation on the contig assembly

- Yes, you can mask the genome and predict gene structure.

get allelic table by blastn to the progenitor diploid genomes speceies (If have two progenitor diploid genomes , do I need align to the different genome to get two allleic table? )

- Yes, you get allelic table using blast based method. One reference genome with chromosomal level assembly should be good enough.

[baozg]

Hi @tangerzhang ，

Thanks for the reply.

1. canu-poly is not recommened for the >2G genome. So what other parameters you recommend? Or would you mind tell what parameters sugar cane genome use ?
2. purge_haplotigs only suit for the diploidy genome. So after canu assembly, I can directly use HiC data？

Best,
Zhigui

[tangerzhang]

Hi Zhigu,
Below are the parameters that I used to assemble a polyploid genome:

canu -s spec.txt -p AP -d run1 genomeSize=3.2g -pacbio-raw input.fasta

Information in spec.txt file:

gridOptions="-N canu -l walltime=30000:00:00 -q all.q"
corOutCoverage=100
ovbMemory=8g
maxMemory=500g
maxThreads=48
ovsMemory=8-500g
ovsThreads=4
oeaMemory=32g
ovsMethod=parallel

For your second question, please use Hi-C data directly once you finish CANU assembly and Pilon polish.