NO haplotype output for HiFi+ Hi-C

[mahajrod]

Hi,

I am playing with verkko2 now. As a test case, I tried to assemble diploid baker yeast genome from HiFi + Hi-C, but got no phased output.
Verkko (v2.3.2) finished without errors. Verkko was installed from conda.
Any idea what have gone wrong? Is Nanopore data a must for phased output?

Sergei

Verkko command:

verkko -d results/verkko/ \
       --hifi results/data/hifi/final/SRR13577846.half.fastq.gz \
       --hic1 results/data/hic/final/SRR7126301.200x__1.fastq.gz \
       --hic2 results/data/hic/final/SRR7126301.200x__2.fastq.gz \
       --local --local-cpus 16

Output files:

0-correction/
1-buildGraph/
2-processGraph/
5-untip/
6-layoutContigs/
7-consensus/
8-hicPipeline/
assembly.colors.csv
assembly.cram
assembly.disconnected.fasta
assembly.fasta
assembly.hifi-coverage.csv
assembly.homopolymer-compressed.gfa
assembly.homopolymer-compressed.layout
assembly.homopolymer-compressed.noseq.gfa
assembly.ont-coverage.csv
assembly.paths.tsv
assembly.scfmap
assembly.unassigned.fasta
emptyfile
hifi-corrected.fasta.gz
snakemake.sh
verkko.yml

[Dmitry-Antipov]

Hi,
HiFi + Hi-C pipeline was not a target for verkko development, but it still should do phasing.
can you copy first ~20 lines of assembly.scfmap and assembly.paths.tsv?

[mahajrod]

Files are small, so I have attached them as well.

head -n 20 assembly.scfmap

path haplotype1-0000001 haplotype1_from_utig4-18
piece000001
end
path haplotype1-0000002 haplotype1_from_utig4-26
piece000002
end
path haplotype1-0000003 haplotype1_unused_utig4-11
piece000003
end
path haplotype1-0000004 haplotype1_unused_utig4-15
piece000004
end
path haplotype1-0000005 haplotype1_unused_utig4-16
piece000005
end
path haplotype1-0000006 haplotype1_unused_utig4-17
piece000006
[N16113N]
piece000007
end

head -n 20 assembly.paths.tsv

name	path	assignment
haplotype1_from_utig4-18	utig4-18+,utig4-34+	HAPLOTYPE1
haplotype1_from_utig4-26	utig4-26+,utig4-10-,utig4-7-,utig4-4-	HAPLOTYPE1
haplotype1_unused_utig4-11	utig4-11+	HAPLOTYPE1
haplotype1_unused_utig4-15	utig4-15+	HAPLOTYPE1
haplotype1_unused_utig4-16	utig4-16+	HAPLOTYPE1
haplotype1_unused_utig4-17	utig4-17-,[N10742N:scaffold],utig4-14-	HAPLOTYPE1
haplotype1_unused_utig4-31	utig4-31+	HAPLOTYPE1
haplotype1_unused_utig4-32	utig4-32+	HAPLOTYPE1
haplotype1_unused_utig4-33	utig4-33+	HAPLOTYPE1
haplotype1_unused_utig4-35	utig4-35+	HAPLOTYPE1
haplotype1_unused_utig4-5	utig4-5+	HAPLOTYPE1
haplotype1_unused_utig4-6	utig4-6+	HAPLOTYPE1
haplotype1_unused_utig4-8	utig4-8+	HAPLOTYPE1
haplotype1_unused_utig4-9	utig4-9+	HAPLOTYPE1
na_unused_utig4-0	utig4-0+	NA
na_unused_utig4-1	utig4-1+	NA
na_unused_utig4-12	utig4-12+	NA
na_unused_utig4-13	utig4-13+	NA
na_unused_utig4-19	utig4-19+	NA

assembly.paths.tsv.txt
assembly.scfmap.txt

[Dmitry-Antipov]

Can you also upload assembly.homopolymer-compressed.noseq.gfa ?
Thank you

[mahajrod]

Here it is

assembly.homopolymer-compressed.noseq.gfa.txt

Full run is available from https://sid.erda.dk/cgi-sid/ls.py?share_id=EipyrYcIrx
It is a 1.7 Gb .tar.gz archive at UCPH ERDA storage

[ksenia-krasheninnikova]

Hello, just to confirm
SRR13577846 and SRR7126301 are the same individual?

[mahajrod]

Hi,

this data was used in VGP's manual for their pipeline some time ago, so I had no doubt that HiFi and Hi-C are from the same individual and never checked it. But I have done this.

And the data is from different strains...
So the unexpected is expected in the results.

Sorry, it is my fault.
We can close this issue....

[Dmitry-Antipov]

There's another problem. According to the assembly size and the graph structure I really believe that corresponding HiFi dataset was haploid (or extremely homozygous > 99.95% IDY). So even proper Hi-C data will not phase this assembly graph since there's nothing to phase.

[mahajrod]

It should be diploid, but indeed it can be extremely homozygous. It seems to be a terrible test dataset. I need to find a different one.
I have started assembly of two highly heterozygous fish species (400 and 800 Mbp) from a similar dataset. 800 Mbp genome has very uneven coverage and seems to be difficult for HiFi sequencing... Let's see...

Thank you!