Draft genome assembly and transcriptome sequencing of the golden algae Hydrurus foetidus (Chrysophyceae)

Culturing of H. foetidus

The specimen of Hydrurus foetidus (Villars) Trevisan (strain G070301) used in this study was sampled from the river at the Finse Alpine Research Center (60°36' N. 07°30' E) in March 2007 and is currently kept in culture at University of Oslo. The photosynthetic H. foetidus was cultured at 4°C with a 14:10 hour light/dark cycle and kept in an adapted Guillard & Lorensen’s WC (Wright’s Chu) medium (Guillard & Lorenzen, 1972) as described by Klaveness & Lindstrøm (2011). To prepare for DNA isolation, the growth of large thalli was promoted by repeated transfer of individual thalli into fresh culture media. Large thalli (0.5–1.0 g wet frozen weight) were collected by removal from the culture medium and immediate transfer to -80°C and storage until further processing. The culture will be deposited in a special culture collection, at the Fraunhofer Culture Collection of Cryophilic Algae (CCCryo).

Illumina sequencing of genomic DNA

Isolation of genomic DNA for Illumina sequencing was performed as part of an unpublished master project (Mehrotra, 2018). Briefly, six individual thalli were used for the DNA isolation. DNA isolation was performed using the DNeasy Plant Mini Kit from Qiagen (Qiagen Inc., Valencia, CA, US). To ensure efficient lysis and homogenization of the external polysaccharide sheath, a few titanium beads were added to the frozen samples and the tubes were shaken using TissueLyser II machine (Qiagen Inc., Valencia, CA, US) for four minutes. After the addition of the lysis buffer and the RNase, tubes were placed in a thermomixer set at 65°C and 800 rpm with 20 second intervals for 30 mins. After adding Buffer AP2 to the lysate, the incubation was done on ice for 15 minutes to allow for better precipitation of the polysaccharides. Further, the extraction kit protocol was followed as is, until the second elution step. Here we reused the flow through from the previous step elution to avoid excessive dilution of the samples. Afterwards, the samples were de-salted and concentrated by ethanol precipitation and resuspension in 100 μl of Milli Q water. Finally, the samples were concentrated even further by pooling all the samples and freeze drying with a Leybold-Heraeus Lyovac GT2 (Leybold-Heraeus, Köln, Germany). The final sample had a concentration of 104 ng/μl and 260/280 ratio of 1.85 and 260/230 ratio of 1.47 as measured on a Nanodrop ND-1000 (ThermoFisher, MA, US).

The isolated and freeze-dried genomic DNA was sent to the Norwegian Sequencing Center (NSC) at the University of Oslo for library preparation and sequencing. Briefly, the sequencing library was made with the Illumina Truseq LT DNA kit (following the rapid mode protocol), with 600–700 bp fragment size and sequenced on two lanes of Illumina HiSeq 2500 with 250 bp paired-end reads (Table 1).

Nanopore sequencing of genomic DNA

Genomic DNA was isolated from two thalli as described above, except that tissue lysis was done using MagNA Lyser Green Beads (Roche, Penzberg, Germany) and shaken for 15 sec at 4 m/sec and the incubation at 65°C was done for 10 min. In addition, the supernatant (after adding buffer AP2) was run through QiaShredder columns to further homogenize the lysate. The DNA was eluted (twice, but re-using the elution buffer) in 50 µl AE elution buffer. To further clean and concentrate the samples, the samples were pooled and cleaned using the Zymo DNA Clean & Concentrate kit (Zymo Research, CA, US). The sample was double-eluted (as before) in 50 µl kit provided elution buffer (DNA concentration 61.8 ng/μl, 260/280 ratio of 1.93 and 260/230 ratio 2.05 as measured on a Nanodrop).

DNA sequencing was done using the MinION (MIN-101B) sequencer, the R9.5 Flow Cell and following the SQK-LSK108 protocol (version GDE_9002_v108_revT_18Oct2016) (Oxford Nanopore, Oxford, UK). Approximately 1 µg of starting DNA was used and inspection of the DNA on a 0.7% agarose gel run at 30 volts from 18 hours showed that the majority of the DNA was between 20-30 kbp, but with a long tail of shorter fragments. The sequencing was run using the MinKNOW software (Oxford Nanopore, Oxford, UK; downloaded October 2017) on an iMac and stopped after 36 hours. Base-calling of the raw Nanopore sequence data was done using Albacore v.2.1.10 (Linux, Python 3.5 version; Oxford Nanopore, Oxford, UK) with default settings. The process was run on the Abel computing cluster at the University of Oslo.

PacBio transcriptome sequencing

Total RNA was isolated from one frozen thallus using Qiagen RNeasy Plant kit, including a QiaShredder column and lysis using MagNA lyser beads as described above, otherwise following the kit protocol. Isolated RNA was sent to NSC for library preparation and PacBio sequencing. Three size fractions (1-2 kbp, 2-3 kbp and 3-5 kbp) were prepared using the IsoSeq library preparation protocol (with selection of polyadenylated transcripts) and sequenced on RSII SMRT cells (Pacific Biosciences, CA, US) (Table 1 and Table 2).

Illumina small RNA sequencing

Small RNAs (below 200 bp) were isolated from a frozen thallus using the Sigma mirPremier kit (Sigma-Aldrich, MO, US) following the manufacturer’s instructions, but including lysis with MagNA beads as described above. The sample was sent to the NSC for library preparation and sequencing. Sequencing library (up to approx. 40 nt fragment size) was prepared and sequenced with Illumina NextSeq 500 as single-end 75 bp reads (Table 1).

Draft genome assembly

The basecalled Nanopore reads were processed with Porechop v0.2.2 using default parameters to remove sequencing adapters. Next, the reads were filtered with Nanofilt v2.0.0 (De Coster et al., 2018) to remove reads shorter than 500 bp and average quality below 9. The filtered reads were further error-corrected with LoRDEC v0.7 (Salmela & Rivals, 2014) using the Illumina reads. First the Illumina reads were quality assessed by removing sequencing adapters and bases with an average quality below 20 (average score across 4 bases), in addition to leading and trailing bases with a quality below 20. This was done using Trimmomatic v0.36 (Bolger et al., 2014). Then lordec-correct (options -k 21 -s 3) was run with the trimmed Illumina reads to correct the filtered Nanopore reads. Then the corrected reads were run through Canu v1.6 (Koren et al., 2017) for further correction (canu -correct with genome Size set to 300 m) and trimming (canu -trim) before assembly (canu -assemble). The assembly was done with two different corrected error rates, 0.144 and 0.146. The two assemblies were almost identical, but the results from using the corrected error rate of 0.144 were used further because the total size was slightly larger and also had the largest contig. The Canu assembly was then polished using the trimmed Illumina reads (described above) by running three rounds of Pilon v1.22 (Walker et al., 2014). The final genome assembly consisted of 5069 contigs with a total length of 171 183 409 nt. The N50 was 43,856 nt and the longest contig of 5,118,963 nt (Table 3).