Is the final assembly file generated?

[hungweichen0327]

Dear developer:

Recently, I ran NovoLoci, but I did not get the final assembly file. Is "assembly_raw_FS003_haploid.fasta" the final assembly file?

This is the content of the config file (FS003_diploid_config.txt):

Project:
-----------------------
Project name          = FS003_haploid
Assembly length       = WG
Genome size           =
Ploidy                = 1
Circular              =
Threads               = 32
Output path           = /home/hwc/nas2_hwc/CPL_Cladosporium/assembly/FS003/NOVOLoci/haploid/
TMP path              = /home/hwc/nas2_hwc/CPL_Cladosporium/assembly/FS003/NOVOLoci/haploid_tmp/

Nanopore reads:
-----------------------
Nanopore reads        = /home/hwc/nas2_hwc/CPL_Cladosporium/ONT_data/merged_barcode06.fastq.gz
Local DB and NP reads =
Sequencing depth NP   = 13
R10                   = yes
Min read length NP    = 1000

PacBio reads:
-----------------------
PacBio reads          =
Local DB and PB reads =
Sequencing depth PB   =
Min read length PB    =

This is the log file of the NovoLoci.
novoloci_FS003_haploid.log.txt

I would like to ask how this issue can be addressed. Thank you very much.

[ndierckx]

Hi,
For the moment the assembly is still very slow, so I think it never finished, what is the time limit you gave?
As reference, if use the WG mode on a 110 Mb genome, it took 4 days. It doesn't take any memory, so you can also run it on your laptop in the background with 10 cores or so. But since you have the newest R10 data and the genome is quite heterozygous, which means phasing should not be too hard, you can probably have a good assembly with Hifiasm in no time? I only use the WG mode of NOVOLoci when I have R9, because there is no good alternative for older ONT data. Or when you really want accurate phasing, because I noticed Hifiasm does create chimeras of both haplotypes.

And I fixed the issue with diploid option in WG mode, but making some additional improvements, should be able to upload it tomorrow.

[hungweichen0327]

Hi @ndierckx,
Thank you for the updates.

This is our own local server, so there is no time limit. However, I am not sure why the final assembly file is not being generated. I also tried running the assembler with another dataset and observed a similar situation (attached is the log file). Do you have suggestions how to solve this problem?
novoloci_NF006_haploid.log.txt

I am currently assembling a Cladosporium species (fungus) genome of about 35 Mb. Since I don’t have Illumina data, I am not sure about the heterozygosity level of this genome. I have assembled four genomes using Hifiasm; three of them gave good assemblies, but one turned out with relatively poor quality (contig number: 1,547, contig N50: 28,122), even though I had about 35× coverage (1,217 Mb, read N50: 2,780). That is why I would like to try NOVOLoci to see if it can produce a better result.

Thank you for the help.

[ndierckx]

1547 contigs is indeed surprisingly bad, NOVOLoci can give some amazing improvements on R9, but on R10+Hifiasm the improvements are more in phasing accuracy instead of contig N50, because Hifiasm is really focused on high N50 values. But I am curious why so I can give a try for you if you want. Could you share the dataset? Because it will be easier to troubleshoot.. From the log I can't say why it terminated, but if I run it myself I probably can

[hungweichen0327]

Hi @ndierckx,

Thank you for the kind help.

This is the Google Drive link to the raw fastq data.
https://drive.google.com/file/d/1nW604VcTKPO95N_amNg-GWCBbRSVM_31/view?usp=sharing

[hungweichen0327]

Hi @ndierckx,

Is there any update? Thank you.

[ndierckx]

Hi,
Sorry for the delay, I had to make an update to make it run better on your data, I started the assembly again today, will let you know when finished.
But the problem was the very short reads, average below 2000 bp and there seem to be some sequencing depth drops that terminate the assembly, so I guess it is normal that Hifiasm didn't perform well..

I will let you know how the assembly went and will upload a new version of NOVOLoci because fixed quite a bit of bugs and made it work better for shorter read lengths. What was the N50 of the Hifiasm assembly?

Greets

[ndierckx]

The assembly was not great either, also 1,500 contigs, think it will be hard to get anything better then you got before
I added the assembly run in the attachment, maybe some of the larger contigs can be used

assembly_final_FS003_diploid_2.fasta.gz

[ndierckx]

Ow it says diploid in the config but just tried haploid:

Project:
-----------------------
Project name          = FS003_diploid_2
Assembly length       = WG
Save assembled reads  = yes
Seed Input            = 
Genome size           = 
Ploidy                = 1
Circular              = 
Threads               = 20
Output path           = /FS003_diploid_2/
TMP path              = /FS003_diploid_2/TMP/

Nanopore reads:
-----------------------
Nanopore reads        = 
Local DB and NP reads = FS003_diploid/
Sequencing depth NP   = 15
R10                   = yes
Min read length NP    = 800

PacBio reads:
-----------------------
PacBio reads          = 
Local DB and PB reads = 
Sequencing depth PB   = 
Min read length PB    = 

[hungweichen0327]

Many thanks for your reply and test.

The output of assembly using Hifiasm:
No. of sequence- 1,547
Contig N50- 28,122

Do you have any suggestions or comments of how to improve this assembly? So far, I had about 35× coverage (1,217 Mb, read N50: 2,780).