Rainer Lehtonen Ph D Genomics and genetics project

Rainer Lehtonen Ph. D, Genomics and genetics project leader Metapopulation Research Group Department of Biological and Environmental Sciences, University of Helsinki Glanville fritillary butterfly genomics and genetics

Glanville fritillary butterfly – genomics and genetics �Background �Genome project �Genome assembly >> Panu Somervuo �Some NGS applications �Conclusions 2

Glanville fritillary as a model �Glanville fritillary is an internationally recognized metapopulation model system in ecological and evolutionary studies �Studied since 1991 in the Åland Islands in Finland �Data available from different populations: - Fragmented landscape vs. continuous Isolated vs. metapopulation Large vs. small Same vs. different population history Field studies, indoor & outdoor cage + laboratory experiments, controlled crosses, molecular studies 3

Collaborative genome project DNA (+RNA) SAMPLES INSTITUTE OF BIOTECHNOLOGY SEQUENCE DATA PRODUCTION INSTITUTE OF BIOTECH, KAROLINSKA INSTITUTE QC + ASSEMBLY INSTITUTE OF BIOTECH, DEP COMPUTER SCI ASSEMBLY VALIDATION (ref g) INSTITUTE OF BIOTECH, DEP COMPUTER SCI ANNOTATION + PUBLICATION EBI, ENSEMBL GENOMES GENOME ANALYSIS EBI, OTHER GENOME PROJECTS VARIATION IN THE GENOME INSTITUTE OF BIOTECH, DEP COMPUTER SCI GENETIC TOOLS FIMM, BIOMEDICUM HKI, INSTITUTE OF BIOTECH, ILLUMINA INC. 4

Reference genome + variation NEX-GEN SEQUENCING 454, SOLi. D 3, SOLEXA REF DNA +RNA SAMPLES EST ASSEMBLY ESTs GENOME ASSEMBLY REF GENOME NEX-GEN RE-SEQUENCING SOLi. D 4/SOLEXA CROSSES/POP POOLS/INDS MAPPING TO REF GENOME VARIATION GENETIC MAP (MARKER LOCATIONS) GENETIC VARIATION GENE EXPRESSION GENOME ANNOTATION DATA FROM OTHER SOURCES PLATFORM FOR LARGE SCALE TARGETED GENOTYPING OF LARGE POPULATION SAMPLES (>50 K)

Variation & other nex-gen data Sample Aim Platform Read Type Read Length Runs to be done RNA, pool used in RNAseq Gene start sites Gene 5’ variation SOLi. D 4 Pair-end 50+25 1/4 Amp DNA, 4 crosses Construction of genetic map SOLi. D 4 Single read, RAD tag library 50+25 3 Amp DNA, pool ~30 ind SNPs & other genetic variation SOLi. D 4 Pair-end 50+25 1 RNA, pooled Variation in 5+2 pop samples pop from 5+1 pop SNPs in ESTs, Expression SOLi. D 4 Pair-end 50+25 1(-2) Single read 400 1/4 DNA from selected individuals 25. -26. 3. 2010 Pgi & flanking genes + Sdhd, Hsp 70 Sure. Select + 454 Heliconius Genome Meeting Sanger seq 6

Deep re-sequencing RAD-tag (Restriction Enzyme Associated DNA) known also as “Deep sequencing of reduced representation library” Example: Construction of a high-density genetic map: *4 controlled Spain-Finland crosses * Parents and 50 individuals from each family to be sequenced Genetic or linkage map defines an order and distance between markers based on a recombination frequency (1 c. M = 1% recombination rate) in meiosis Sure. Select (Agilent)Target Enrichment + deep sequencing with 454 Example: Population comparison of the Pgi + flanking genes (+ some other) in a sample of 24 individuals or pools 7

Genetic map with RAD-tag NGS Nathan A et al. Plo. S ONE 2008 Now: 500 M Reads 50 bp each 150 -200 bp pair-end library 50 bp seq SNP 1 25 bp seq SNP 2 8

RAD-tagging in Glanville fritillary Average fragment size 454 Glanville g. Contigs Nco. I 13. 3 Xho. I 11. 5 Eco. RI 4. 5 Heliconius 14 4 2 Mappable reads • Restriction site > 250 bp from the end of a g. Contig • Targets = 2 x sites • 454 -Newbler assembly: 320 Mbp (out of ~550 Mbp genome in 220 K contigs (>500 bp) • Expected number of SNPs 1/300 bp, read lenght 50 -25 bp --------------------------- #sites #mappable #exp #SNPs Nco. I* ccatgg 24, 064 38, 880 48, 128 12, 032 Xho. I ctcgag 27, 788 45, 925 55, 576 13, 894 Eco. RI gaattc 70, 474 117, 293 140, 948 35, 2367 Bsph. I* tcatga 66, 967 110, 731 133, 934 33, 483 Nde. I catatg 73, 629 121, 628 147, 258 36, 814 *The most probable combination > ~45, 000 SNPs • Reads have to unique • 10 -20 x coverage/ individual (>~5000 x on average) • Heavy data filtering needed > probably only 30 -50% of data is usable In silico restriction analysis made by Panu Somervuo, MRG 9

Targeted enrichment + resequencing Max 55 K 120 mer oligos Glanville fritillary butterfly Sure. Select Target enrichment (10 x tiling): • To identify “lethal” haplotypes associated to a known homozygous genotype • To define structure and variations of the hypervariable Pgi gene * To design tag-SNPs for large scale genotyping 10

Uneven coverage Hypothesis driven sampling compare samples (24) from different populations with different tag-SNP genotype frequencies >Hardy-Weinberg equilibrium > Hardy-Weinberg disequilibrium Cinxia Sure Select 5468 TCMID_72 - Tas_pooli_Cinxia Sure Select_13 -16 TCMID 71 - TCMID 70 - TCMID_69 - Tas_pooli_Cinxia Sure Select_E 3 TCMID_68 - Tas_pooli_Cinxia Sure Select_D 3 TCMID_67 - Tas_pooli_Cinxia Sure Select_5 TCMID_66 - Tas_pooli_Cinxia Sure Select_4 TCMID_65 - Tas_pooli_Cinxia Sure Select_3 TCMID_64 - Tas_pooli_Cinxia Sure Select_2 TCMID_63 - Tas_pooli_Cinxia Sure Select_1 TCMID_62 - Tas_pooli_Cinxia Sure Select_C 3 TCMID_61 - Tas_pooli_Cinxia Sure Select_B 3 TCMID_60 - Tas_pooli_Cinxia Sure Select_A 3 TCMID_59 - Tas_pooli_Cinxia Sure Select_A 2 TCMID_58 - Tas_pooli_Cinxia Sure Select_H 1 TCMID_57 - Tas_pooli_Cinxia Sure Select_G 1 TCMID_56 - Tas_pooli_Cinxia Sure Select_F 1 TCMID_55 - Tas_pooli_Cinxia Sure Select_E 1 TCMID_54 - Tas_pooli_Cinxia Sure Select_D 1 TCMID_53 - Tas_pooli_Cinxia Sure Select_C 1 TCMID_52 - Tas_pooli_Cinxia Sure Select_B 1 TCMID_51 - Tas_pooli_Cinxia Sure Select_A 1 TCMID_50 - Tas_pooli_Cinxia Sure Select_6, 9 -12 +7+8 TCMID_3 - Tas_pooli_Cinxia Sure Select_F 3 14 731 7774 7960 6324 7718 3708 3621 6499 5361 4983 3613 4494 21 122 22 316 17 110 20 851 9 780 9 214 16 644 13 717 12 959 9 362 3128 5 000 Reads (total 337 635) 9 863 7 520 9 164 10 540 5236 4343 Bases kbp (total 128 555 kbp) 11 687 444 1 131 3581 2829 3587 1791 4 568 4144 � • 13 346 8204 20 699 11 072 7 998 12197 11546 10 000 15 000 31 488 30 753 20 000 25 000 30 000 35 000 ¼ 454 Titanium run: 444 -12197 kb/sample = 15 -406 x coverage Figure by Pia Laine Institute of Biotechnology University of Helsinki 11

How well Sure. Select works? Our very preliminary result: ~40% of the data comes from the target Data from Agilent 12

Comparison of haplotypes Sampsa Hautaniemi, Marko Laakso, Sirkku Karinen, Rainer Lehtonen Sirkku. Karinen@helsinki. fi 25. -26. 3. 2010 Heliconius Genome Meeting 13

Message �Whole genome sequencing is doable for a “non-genome” oriented research group �Most work on data filtering and analysis �Tools for data management and analysis under strong development �Down-stream efforts need to be compatible with available genome data 14