DNA sequencing data

DNA sequencing approaches and data

Per Unneberg

Sequencing technologies

Illumina NovaSeq 600

Scale up and down with a tunable output of up to 6 Tb and 20B single reads in < 2 days.

Up to 2X250 bp read length. Price example: 8,000 SEK total for resequencing 3Gbp genome to 30X

https://www.illumina.com/systems/sequencing-platforms/novaseq.html

PacBio Revio

Up to 360 Gb of HiFi reads per day, equivalent to 1,300 human whole genomes per year.

Tens of kilobases long HiFi reads. Price example (Sequel II): ~35kSEK per library and SMRT cell

https://www.pacb.com/revio/

Sequencing approaches

Figure 1: DNA sequencing costs (Wetterstrand, KA)

Despite price drop, still need to make choices regarding depth and breadth of sequencing coverage and number of samples.

Figure 2: Comparison of common sequencing approaches. Restriction site-associated DNA sequencing (RAD-seq, top), targets regions flanking given restriction sites, but misses much of genome. Pooled sequencing (Pool-seq, middle) is cost-effective, but loses information about individuals. Low-coverage whole genome sequencing (lcWGS, bottom) is increasing in popularity, but genotyping low coverage data is problematic.

Lou et al. (2021)

Our focus will be on Whole Genome reSequencing (WGS), mostly high-coverage.

Despite the fact that sequencing costs have dropped dramatically (left), there still are choices to be made regarding the distribution of costs along 1) sequencing coverage depth, i.e., the mean depth of sequencing 2) sequencing coverage breadth, i.e., whether or not to do targeted or whole-genome resequencing or 3) sample size; how many individuals to sample. Whole-genome resequencing of individuals from populations to sufficient depth (30X) is still very expensive, but often needed to understand mechanisms of adaptation (Lou et al., 2021, p. 5967). Various protocols have been developed to meet the challenges that cost imposes:

1. RAD-seq, restriction site-associated DNA sequencing, targets regions flanking given restriction sites. Downside: much of genome is missed
2. pool-seq, pooled sequencing. Cost-effective, but loses information about individuals
3. lcWGS, low-coverage whole genome sequencing increasing in popularity. Genotyping low coverage is problematic however.

Our focus here is WGS (whole-genome resequencing), primarily high-coverage, despite the cost it may incur.

Genome assembly and population resequencing

Genome assembly

Figure 3: Schematic overview of genome assembly and annotation. Population resequencing often requires a reference assembly. Assembly quality impacts downstream analyses

Allendorf et al. (2022)

Population resequencing

Figure 4: Overview of short reads mapped to a reference sequence. Reads have been colored according to population (red or yellow). There are three individuals from each population. Note the variation in sequence coverage. Mismatches in reads are highlighted as different colors with respect to population color.

DNA sequences are generated from DNA fragments, often as paired-end reads, or long reads. For population genomics, it is often necessary to generate a reference sequence through genome assembly, the quality of which will impact the types of downstream analyses that can be done. Reads are assembled into contigs (contiguous sequences) that with the aid of genetic maps are pieced together into scaffolds. In the left figure, note that the scaffolds contain areas of unknown sequence.

Once a reference sequence has been produced, (short) reads from individuals are mapped to the reference (right).

DNA sequences in FASTQ format

ls --long --human fastq/*.fastq.gz

-rw-r--r-- 1 runner runner 302K Sep 18 22:10 fastq/PUN-Y-INJ_R1.fastq.gz
-rw-r--r-- 1 runner runner 393K Sep 18 22:10 fastq/PUN-Y-INJ_R2.fastq.gz

Count number of lines:

zcat fastq/PUN-Y-INJ_R1.fastq.gz | wc --lines

15768

Format:

1. sequence id (prefixed by @)
2. DNA sequence
3. separator (+)
4. Phred base quality scores

zcat fastq/PUN-Y-INJ_R1.fastq.gz | head --lines 8 | cut --characters -30

@SRR9309790.10003134
TAAATCGATTCGTTTTTGCTATCTTCGTCT
+
AAFFFJJJJJJJFJJJJJJJJJJJJJJJJJ
@SRR9309790.10003222
TAAATCGATTCGTTTTTGCTATCTTCGTCT
+
AAFFFJJJJJJJJJJJJJJJJJJJJJJJJJ

First data encounter is usually FASTQ files from sequencing center. Files contain sequence reads that are generated by the sequencing machine. They correspond to short stretches of DNA that have been sequentially read by the machine, resulting in a string of base calls that constitute the reads. Every DNA base has an associated so-called Phred-scaled quality score from 0 to 93 that are encoded using ASCII 33 to 126.

The format consists of 1. @ title with id 2. raw sequence 3. +: a separator with optional description 4. the quality scores

DNA sequence quality control

Quality values represent the probability P that the call is incorrect. They are coded as Phred quality scores Q. Here, Q=20 implies 1% probability of error, Q=30 implies 0.1% and so on. Typically you should not rely on quality values below 20.

Q = -10 \log_{10} P

A common way to do QC is with fastqc:

fastqc --outdir fastqc --extract fastq/*fastq.gz

• For Illumina paired-end sequencing, the second read pair usually shows a larger drop in quality towards ends
• Trimming the sequences for adapter sequence and quality is good practice (e.g., with CutAdapt (Martin, 2011))

Bibliography

Allendorf, F. W., Funk, W. C., Aitken, S. N., Byrne, M., & Luikart, G. (2022). Population Genomics. In F. W. Allendorf, W. C. Funk, S. N. Aitken, M. Byrne, G. Luikart, & A. Antunes (Eds.), Conservation and the Genomics of Populations (p. 0). Oxford University Press. https://doi.org/10.1093/oso/9780198856566.003.0004

Lou, R. N., Jacobs, A., Wilder, A. P., & Therkildsen, N. O. (2021). A beginner’s guide to low-coverage whole genome sequencing for population genomics. Molecular Ecology, 30(23), 5966–5993. https://doi.org/10.1111/mec.16077

Martin, M. (2011). Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet.journal, 17(1), pp. 10–12. https://doi.org/10.14806/ej.17.1.200

Wetterstrand, KA. DNA Sequencing Costs: Data from the NHGRI Genome Sequencing Program (GSP). www.genome.gov/sequencingcostsdata