OmicsBox now includes Hifiasm for de novo assembly from PacBio HiFi or Oxford Nanopore R10 reads. It complements the existing ABySS, SPAdes and Flye assemblers.
What is Hifiasm
Hifiasm is a de novo assembler for long, accurate reads developed in Heng Li’s group (Cheng et al., Nature Methods 2021). It was designed for PacBio HiFi reads and now also assembles Oxford Nanopore R10 data.
Diploid and polyploid organisms carry two or more copies of each chromosome. Most assemblers collapse these copies into a single consensus that mixes alleles from different haplotypes. Heterozygous variation is lost, and divergent regions may appear as false duplications. Because HiFi reads have an accuracy above 99%, the differences between haplotypes can be read directly from the data, which makes haplotype-resolved assembly possible.
Hifiasm corrects sequencing errors while preserving true heterozygous differences, then builds a phased assembly graph in which haplotype-specific sequences remain separate. Additional data extend phasing beyond single reads: short reads from both parents (trio binning) or Hi-C reads link phased blocks across entire chromosomes.
Main advantages:
- Haplotype-resolved assemblies instead of a mosaic consensus.
- Speed and efficiency: vertebrate-sized genomes can be assembled on a single server.
- Polyploid genomes supported.
- Widely adopted by large reference projects such as the Human Pangenome Reference Consortium and the Darwin Tree of Life.
Hifiasm is most useful for heterozygous species, such as outbred animals and wild plants, and for studies that require both haplotypes, for example allele-specific expression or structural variation between haplotypes. With trio or Hi-C data, it produces chromosome-scale phased assemblies. For short-read data, the ABySS and SPAdes assemblers in OmicsBox remain the appropriate choice.
How to run Hifiasm in OmicsBox
Hifiasm is available in OmicsBox, a desktop application for NGS data analysis that requires no programming. Input reads, phasing strategy and parameters are selected in its graphical interface, and Hifiasm does not need to be installed.
The tool is located under Genome Analysis → DNA-Seq de Novo Assembly → Hifiasm. To run an assembly you need:
- Long reads: PacBio HiFi or Nanopore R10 simplex reads (one technology per run). Nanopore ultra-long reads can be added to improve contiguity.
- Phasing data (optional): short reads from both parents for Trio Binning, or paired-end Hi-C reads for Hi-C Phasing. Without them, Hifiasm produces a primary assembly and two partially phased haplotypes.
Ploidy, duplication purging and coverage options can be adjusted. For inbred or homozygous samples, a collapsed primary-only assembly is recommended. An optional telomere motif (e.g. CCCTAA in vertebrates) helps identify chromosome ends.
Monitoring the analysis progress
De novo genome assembly requires substantial memory and CPU time. In OmicsBox, Hifiasm runs on the OmicsBox cloud, which provides these high-memory resources ready to use. There is no need to set up or maintain a server or computing cluster, and the job completes faster than on a typical workstation.
With CloudSync, OmicsBox only needs to be open to configure and submit the job. OmicsBox can then be closed and the computer shut down while the assembly continues in the cloud. An email notification can be requested when the job finishes, and results can be saved to Cloud Files to open them from any device signed in with the same account.
With OmicsBox Web, the progress of an assembly can be checked at any time and from anywhere: a laptop, tablet or phone with a browser is enough, with no installation. The Jobs page shows the status, progress, run time and logs of each analysis, and running jobs can be cancelled. Once the assembly is done, results can be previewed directly in the browser and downloaded, so a first look at the assembly statistics does not have to wait until you are back at your desk.
Results
Assemblies are written in standard formats and can be used directly in other tools:
- FASTA files of the primary assembly and the haplotype assemblies. Collapsed runs also return alternate contigs.
- GFA assembly graphs (optional), viewable in Bandage, and BED files of low-quality regions.
Within OmicsBox, a summary report collects the input data and QUAST assembly statistics, including N50 and L50 for each FASTA file. It also includes configurable charts, such as an Nx plot and a k-mer spectrum, that give a quick view of assembly quality.
The Nx plot compares the contiguity of the primary assembly and the haplotypes, while the k-mer spectrum shows the expected heterozygous and homozygous peaks for a diploid sample, and additional peaks may indicate contamination or repeats. All charts can be exported to high-quality png or pdf files for presentations or publications.
Assemblies can then be annotated with EviAnn and functionally annotated with Blast2GO within OmicsBox.
Availability
Hifiasm is available in the Genome Analysis module. All parameters are described in the Hifiasm user manual.
Reference: Cheng H, Concepcion GT, Feng X, Zhang H, Li H. Haplotype-resolved de novo assembly using phased assembly graphs with hifiasm. Nature Methods 18(2), 170–175 (2021).