A Gene-Based Genetic Linkage Map of the Collared Flycatcher (Ficedula albicollis) Reveals Extensive Synteny and Gene-Order Conservation During 100 Million Years of Avian Evolution

Abstract

By taking advantage of a recently developed reference marker set for avian genome analysis we have constructed a gene-based genetic map of the collared flycatcher, an important “ecological model” for studies of life-history evolution, sexual selection, speciation, and quantitative genetics. A pedigree of 322 birds from a natural population was genotyped for 384 single nucleotide polymorphisms (SNPs) from 170 protein-coding genes and 71 microsatellites. Altogether, 147 gene markers and 64 microsatellites form 33 linkage groups with a total genetic distance of 1787 cM. Male recombination rates are, on average, 22% higher than female rates (total distance 1982 vs. 1627 cM). The ability to anchor the collared flycatcher map with the chicken genome via the gene-based SNPs revealed an extraordinary degree of both synteny and gene-order conservation during avian evolution. The great majority of chicken chromosomes correspond to a single linkage group in collared flycatchers, with only a few cases of inter- and intrachromosomal rearrangements. The rate of chromosomal diversification, fissions/fusions, and inversions combined is thus considerably lower in birds (0.05/MY) than in mammals (0.6–2.0/MY). A dearth of repeat elements, known to promote chromosomal breakage, in avian genomes may contribute to their stability. The degree of genome stability is likely to have important consequences for general evolutionary patterns and may explain, for example, the comparatively slow rate by which genetic incompatibility among lineages of birds evolves.

GENOMICS is in a phase where new technology allows genome characterization beyond that of traditional model organisms and species of medical or agricultural interest. For example, genomic analyses of nonmodel species holds great promise for dissecting the genetic background to fitness traits in natural populations, to adaptive population divergence, to speciation, and to other key aspects of evolutionary biology (Ellegren and Sheldon 2008). Genomic characterization of new and phylogenetically divergent lineages has the additional benefit that it provides the necessary comparative perspective for addressing the evolution of genome organization. Specifically, with genetic maps or genome sequence information available across taxa, the broad-scale pattern of genome and chromosomal evolution can be investigated. This, in turn, opens the possibility of investigating to what extent evolution at the chromosomal level sets the stage for the evolutionary processes, which occur on the level of the phenotype.

Reshuffling of chromosomal segments, through translocations and inversions, is an integral part of genome evolution. However, it is clear that the rate of rearrangement differs radically among lineages as well as on a temporal scale (Kohn  et al. 2006; Ferguson-Smith and Trifonov 2007). From comparative mapping of chicken and different mammals it was suggested that the rate of chromosomal rearrangement in the avian lineage is very low (Burt  et al. 1999). This has subsequently been confirmed through analyses of vertebrate genome sequence data, including chicken (Bourque  et al. 2005), the only bird that has had its genome sequenced to date (International  Chicken  Genome  Sequencing  Consortium 2004). Moreover, evidence for an unusually stable avian karyotype with few interchromosomal rearrangements has been obtained by cross-species chromosome painting or the use of other types of in situ hybridization probes (Shetty  et al. 1999; Shibusawa  et al. 2001, 2004a,b; Raudsepp  et al. 2002; Guttenbach  et al. 2003; Kasai  et al. 2003; Derjusheva  et al. 2004; Schmid  et al. 2005; Itoh  et al. 2006; Fillon  et al. 2007; Griffin  et al. 2007; Nishida-Umehara  et al. 2007). However, these experiments rarely have the resolution for detecting intrachromosomal or small-scale interchromosomal rearrangements.

Genetic maps are available for turkey (Reed  et al. 2005) and quail (Kayang  et al. 2006), two agricultural species that are closely related to chicken as members of the order Galliformes. However, a lack of genetic markers, in particular those informative for comparative mapping, has been a major obstacle to linkage analyses of bird species belonging to other orders. As a consequence, linkage mapping in natural bird populations is still in its infancy. Hansson  et al. (2005) developed a partial microsatellite-based linkage map (58 markers) in the great reed warbler (Acrocephalus arundinaceus), a species from the order Passeriformes, the largest and ecologically most well-studied group of birds. This study also made the unexpected observation that the recombination rate was twice as high in females as in males, which is in contrast to the prevailing trend of recombination usually being lower in the heterogametic sex (in birds, males are ZZ and females ZW). Backström  et al. (2006) reported on a gene-based linkage map of the Z chromosome of another passerine species, the collared flycatcher (Ficedula albicollis). The Z chromosome was found to be completely syntenic between collared flycatcher and chicken. Subsequently, Dawson  et al. (2007) developed an extended great reed warbler map and found a high degree of chromosomal conservation when compared to chicken (see also Åkesson  et al. 2007).

We have recently adopted the comparative anchor-tagged sequences approach (Lyons  et al. 1997) to develop a genomewide, gene-based marker resource for avian comparative mapping (Backström  et al. 2008). This set of 200+ markers target conserved exonic sequences in genes spread over all chromosomes currently covered in the chicken genome assembly, with a mean marker interval of 4 Mb. The uniform distribution of these markers across the chicken genome means that, if they are used for comparative mapping in other birds, the degree of synteny and gene-order conservation across a significant part of the avian genome can be revealed. Here we present a genetic linkage map of the collared flycatcher based on the new marker set. This species has long been in focus for studies of sexual selection, life-history evolution, and speciation (Gustafsson and Sutherland 1988; Gustafsson and Pärt 1990; Gustafsson  et al. 1995; Ellegren  et al. 1996; Qvarnström  et al. 2000, 2006; Veen  et al. 2001; Saether  et al. 2007) and hence is a well-established “ecological model organism.” Importantly, songbirds (passeriforms) and galliforms diverged at the time of the major radiation of avian lineages ≈100 million years (MY) ago (Van  Tuinen  et al. 2000). With genetic map data for the collared flycatcher we can thus address genome evolution at the level of gene order across two highly divergent lineages of the avian phylogenetic tree.

MATERIALS AND METHODS

Species samples and DNA extraction:

Blood samples were collected from collared flycatcher (F. albicollis) families breeding on the Baltic islands Öland and Gotland and DNA was extracted by a standard proteinase K digestion/phenol–chlorophorm purification protocol. The mapping pedigree consisted of 24 half-sib families with a few interconnections and 11 F₂'s, in total 322 birds (supplemental Table 1) after excluding all recognized extra-pair offspring (see below).

Marker genotyping:

In a previous resequencing effort, we surveyed 200 collared flycatcher genes for intronic diversity among 10 unrelated individuals from the same population as the mapping pedigree, which uncovered 904 segregating sites (Backström  et al. 2008). From this, 341 single nucleotide polymorphisms (SNPs) with a minor allele frequency of >0.1 and representing the majority of all genes screened were selected for genotyping in the pedigree; for many genes, more than one SNP from the same intron were included. An additional 43 SNPs were obtained from 21 different genes previously screened for variability in collared flycatchers (Borge  et al. 2005) (Table 1). The total of 384 SNPs were genotyped using the Golden Gate Assay (Fan  et al. 2003) from Illumina (San Diego) at the SNP Technology Platform, Uppsala University (http://www.medsci.uu.se/molmed/snpgenotyping/index.htm). The overall genotype call rate was 95.7% and the reproducibility was 100% according to duplicate analysis of 5.4% (7218/132,848) of the genotypes. The quality of the genotype data was further assessed by testing for Hardy–Weinberg equilibrium (HWE) using the chi-square distribution for each assay. All SNPs conformed to HWE.

TABLE 1

Markers included in the collared flycatcher linkage map

Marker	Linkage group	Gene descriptiona	Ensembl IDb	Chicken chromosome	Chicken genome start position (bp)
Gene-based SNPs
00548	Unlinked	Hypothetical protein	00548	22	1,221,445
02079	Unlinked	Hypothetical protein	02079	10	1,854,152
02419	Fal9	No longer in the Ensembl database, not mapped to new identifiers
04550	Fal27	No longer in the Ensembl database, not mapped to new identifiers
05087	Fal13	Hypothetical protein	05087	19	5,590,343
07726	Fal16	Hypothetical protein	07726	11	10,700,742
08235	Fal6	Hypothetical protein	08235	Un	42,824,332
08544	Fal10	No description	08544	6	17,299,861
12630	Fal28	Magmas-like protein	12630	14	13,386,833
15691	Fal6	Uncharacterized protein C15orf24 precursor	15691	5	32,353,219
15738	Fal5	Similar to CG1218-PA	15738	4	26,412,511
17140	Fal5	No longer in the Ensembl database, not mapped to new identifiers
18798	Fal6	Kinesin light chain	18798	5	52,944,389
20352	Fal1	No description	20352	2	56,437,562
20904	Fal1	Hypothetical protein	20904	2	67,141,279
21277	Fal1	No description	21277	2	80,448,704
22644	Fal2	No description	22644	3	58,129,917
25613	Fal8	Hypothetical protein	25613	1	108,461,452
25924	Unlinked	Hypothetical protein	25924	4	93,260,081
27425	Fal4	Hypothetical protein	27425	1	171,909,504
27623	Fal4	No longer in the Ensembl database, not mapped to new identifiers
ABHD10	Fal8	Abhydrolase domain-containing mitochondrial precursor	24813	1	91,842,859
ACADL	Fal7	Acyl-coenzyme A dehydrogenase, long chain	04557	7	2,734,809
ACADSB	Fal10	Acyl-coenzyme A dehydrogenase, short/branched chain	15724	6	33,024,247
ACHA9	Fal5	Neuronal acetylcholine receptor subunit α-9 precursor	23080	4	70,872,474
ACLYc	Fal1	ATP citrate lyase	05502	27	4,344,212
ACOT8	Fal15	Acyl-coenzyme A thioesterase 8	11074	20	10,473,488
ACTBc	Fal32	Actin, cytoplasmic type 5	39969	10	1,891,946
ADAL	Fal23	Adenosine deaminase-like	06419	10	7,196,900
ADH5	Fal5	Alcohol dehydrogenase 5 (class III), χ polypeptide	19994	4	61,539,229
ADIPOR1	Fal27	Adiponectin receptor 1	00132	26	1,090,425
ALAS1c	Fal21	5-Aminolevulinate synthase, nonspecific, mitochondrial precursor	06295	12	2,762,886
AN32B	Fal26	Acidic leucine-rich nuclear phosphoprotein 32 family member B	02401	28	1,324,041
ANAPC5	Fal9	Anaphase-promoting complex subunit 5	06640	15	5,498,421
ANKRD49	Fal4	Ankyrin repeat domain-containing protein 49	27818	1	189,909,241
ARF1	Unlinked	ADP-ribosylation factor 1	08661	2	2,259,733
ARHGEF9	Unlinked	Rho guanine nucleotide exchange factor 9	12303	4	11,866,519
ARHL2	Unlinked	Poly(ADP-ribose) glycohydrolase ARH3	03624	23	4,457,042
ARP6	Fal3	Actin-related protein 6	18851	1	49,106,063
ASB6	Fal25	Ankyrin repeat and SOCS box-containing 6	06983	17	6,201,193
ATG4B	Fal14	Cysteine protease ATG4B	10179	9	5,828,498
ATP6AP2	Fal4	ATPase, H+ transporting, lysosomal accessory protein 2	26187	1	115,861,428
ATP6V1E1	Unlinked	Vacuolar H+ ATPase E1	21281	1	63,935,883
BZW1	Fal7	Basic leucine zipper and W2 domain-containing protein 1	13380	7	12,346,390
C12orf29	Fal3	Hypothetical protein	18208	1	44,668,513
C7orf27	Fal32	HEAT repeat domain-containing protein C7orf27 precursor	06938	14	3,350,565
C8orf53	Fal1	Uncharacterized protein C8orf53	25969	2	140,964,418
CACYBP	Fal19	Calcyclin-binding protein	07248	8	7,413,974
CATB	Fal2	Cathepsin B precursor	26896	3	110,173,920
CBPZ	Fal5	Carboxypeptidase Z precursor	25149	4	84,149,787
CCDC104	Unlinked	Coiled-coil domain-containing protein 104	13093	3	129,733
CCDC132	Fal1	Coiled-coil domain containing 132	15463	2	22,899,725
CCDC137	Unlinked	MGC16597 protein	07177	18	9,164,974
CCNG1	Fal12	Cyclin-G1	02636	13	6,483,608
CCT2	Fal3	Chaperonin-containing TCP1, subunit 2	16215	1	37,378,125
CDH9	Fal1	Cadherin-9 precursor	21079	2	72,557,369
CEPU1c	Fal24	Protein CEPU-1 precursor	29072	24	1,827,157
CGI-62	Fal1	UPF0418 protein C8orf70	25374	2	125,087,485
CHCc	Fal13	Clathrin heavy chain	39267	19	7,239,507
CHD1L	Fal8	Chromodomain helicase DNA-binding protein 1-like	24254	1	83,862,026
CHM1B	Fal18	Charged multivesicular body protein 1b	06500	4	1,509,856
CHMP5	Fal1	Charged multivesicular body protein 5	21491	2	87,941,041
CNTN1	Fal3	Contactin-1 precursor	15506	1	30,637,524
COEA1	Fal1	Collagen α-1(XIV) chain precursor	26472	2	142,375,665
CRIPT	Fal2	Postsynaptic protein CRIPT	16264	3	27,956,254
CT030	Unlinked	UPF0414 transmembrane protein C20orf30	00227	22	346,000
DC1L1	Fal1	Cytoplasmic dynein 1 light intermediate chain 1	18728	2	40,471,155
DDAH1	Fal11	Dimethylarginine dimethylaminohydrolase 1	14108	8	16,977,584
DECR1	Fal1	2,4-dienoyl-CoA reductase, mitochondrial precursor	25647	2	129,135,915
DLD	Fal3	Dihydrolipoamide dehydrogenase	12884	1	15,844,353
DPYSL3	Unlinked	Dihydropyrimidinase-like 3	12260	13	18,592,762
DST	Fal2	Bullous pemphigoid antigen 1	26267	3	89,753,995
EDF1	Unlinked	Endothelial differentiation-related factor 1 homolog	14657	17	932,044
EF1A1	Fal2	Elongation factor 1-α 1	25653	3	84,252,820
EF1A	Fal15	Elongation factor 1-α	09385	20	8,992,662
EIF3S1	Unlinked	Eukaryotic translation initiation factor 3, subunit 1-α	13336	10	21,935,894
ENO1c	Fal20	α-Enolase	03745	21	3,197,152
ETNK1	Fal3	Ethanolamine kinase 1	21571	1	68,619,958
FAK1	Fal1	Focal adhesion kinase 1	26060	2	151,344,071
FNc	Fal7	Fibronectin	05663	7	4,362,118
FTHc	Fal6	Ferritin H-subunit	11687	5	8,042,629
GAS7	Unlinked	Growth-arrest-specific protein 7	00500	18	181,941
GH1c	Fal8	Growth hormone factor 1	24989	1	96,197,113
GNB1	Fal20	Guanine nucleotide-binding protein (G protein) β polypeptide 1	02040	21	1,907,993
GTF2B	Fal11	Transcription initiation factor IIB	10015	8	15,855,345
HARS	Fal12	Histidyl-tRNA synthetase	01152	13	825,160
HEPACAM	Fal24	Hepatocyte cell adhesion molecule	00574	24	244,611
HMGB2	Fal5	High-mobility group protein B2	17483	4	44,739,576
HMGN2	Fal29	Nonhistone chromosomal protein HMG-17	00504	23	132,497
IGF2R	Fal2	Insulin-like growth factor 2 receptor	18986	3	47,356,791
IGFBP7	Fal5	Insulin-like growth factor-binding protein 7 precursor	18503	4	50,637,829
KCNIP4	Unlinked	Kv channel-interacting protein 4	23272	4	77,264,411
KCRBc	Fal6	Creatine kinase B-type	18765	5	52,833,368
KIAA1706	Fal1	CDNA FLJ14480 fis, clone MAMMA1002215	19789	2	46,847,660
LARP1	Fal12	La-related protein 1	06374	13	12,129,582
LDHA	Fal6	L-lactate dehydrogenase A chain	10181	5	13,644,404
LHCGRc	Fal22	Luteinizing hormone/choriogonadotropin receptor	14806	3	7,517,756
MAGOH	Fal11	Mago-nashi homolog, proliferation-associated	17388	8	25,398,748
MBP	Fal1	Myelin basic protein	22187	2	92,901,749
METRNL	Unlinked	Meteorin-like protein precursor	02154	18	3,058,036
MIC1	Fal1	Colon cancer-associated protein Mic1	24206	2	106,186,277
MITD1	Fal4	MIT domain-containing protein 1	27060	1	136,635,855
MMAA	Fal5	Methylmalonic aciduria type A protein, mitochondrial precursor	16214	4	32,303,140
MOSPD2	Fal4	Motile sperm domain-containing protein 2	26743	1	125,737,284
MPP1c	Fal18	Myelin proteolipid protein	NA	NA	NA
MPP6	Fal1	MAGUK p55 subfamily member 6	17898	2	31,506,073
MRPS18A	Fal2	28S ribosomal protein S18a, mitochondrial precursor	16751	3	32,021,644
NAT5	Fal33	N-acetyltransferase 5	20554	7	38,349,143
NDST3	Fal12	Bifunctional heparan sulfate N-deacetylase/N-sulfotransferase 3	19599	4	56,617,452
NDUFA7	Fal33	NADH dehydrogenase 1 α subcomplex subunit 7	00895	28	871,010
NSMAF	Fal1	Protein FAN (factor associated with N-SMase activation)	24908	2	115,869,042
NY-SAR-48	Fal26	Sarcoma antigen NY-SAR-48 isoform a	05915	28	3,757,065
OAZ	Fal26	Ornithine decarboxylase antizyme	01183	28	1,446,253
ODC11c	Fal2	Ornithine decarboxylase	26527	3	99,660,031
PARK7	Fal20	Protein DJ-1 Parkinson disease protein 7 homolog	00742	21	235,467
PDCD11	Fal10	RRP5 protein homolog programmed cell death protein 11	13446	6	25,047,124
PDHL1	Fal4	Phosphoglycerate dehydrogenase-like 1	27270	1	148,754,342
PES1	Fal9	Pescadillo homolog 1	12619	15	11,118,356
PKHB2	Fal14	Pleckstrin homology domain-containing family B member 2	03399	9	3,184,527
PNN	Fal6	Pinin	16532	5	40,003,168
POLR2C	Fal16	DNA-directed RNA polymerase II subunit RPB3	01768	11	552,555
POLR2H	Fal14	DNA-directed RNA polymerases I, II, and III subunit RPABC3	13907	9	17,035,879
PPIL4	Fal2	Peptidyl-prolyl cis-trans isomerase-like 4	20195	3	50,008,973
PRS4	Fal6	26S protease regulatory subunit 4	17367	5	46,255,522
PSMB1	Fal2	Proteasome (prosome, macropain) subunit, β-type, 1	18217	3	42,604,002
PSMC2	Fal3	Proteasome (prosome, macropain) 26S subunit, ATPase, 2	13403	1	13,964,547
PSMC3	Fal6	Proteasome 26S ATPase subunit 3	13163	5	25,021,822
PSMC5	Fal17	26S protease regulatory subunit 8	00469	27	1,607,679
PSMD14	Fal7	Proteasome (prosome, macropain) 26S subunit, non-ATPase, 14	18142	7	22,963,970
PSMD6	Fal21	26S proteasome non-ATPase regulatory subunit 6	11836	12	13,910,990
PTMS	Fal8	Parathymosin	23363	1	80,288,088
RAB12	Fal1	Gallus gallus similar to Rab12 protein	22528	2	101,592,940
RAB3GAP1	Fal7	Rab3 GTPase-activating protein catalytic subunit	19948	7	32,114,339
RABL4	Fal3	Putative GTP-binding protein RAY-like	20454	1	53,623,713
RASGEF1A	Fal10	RasGEF domain family, member 1A	03978	6	5,681,912
RBBP7	Fal4	Histone-binding protein RBBP7	26698	1	124,845,745
RBM18	Unlinked	Probable RNA-binding protein 18	02045	17	9,445,180
RBM26	Fal4	RNA-binding protein 26	27331	1	157,186,094
RHOc	Fal31	Rhodopsin (opsin-2)	33236	12	20,163,795
RL13	Fal30	60S ribosomal protein L13	09974	11	20,692,991
RNASEH1	Fal2	Ribonuclease H1	26438	3	96,571,133
ROBO1	Fal9	Roundabout1 protein	25008	1	99,689,366
RPL5c	Fal11	60S ribosomal protein L5	09525	8	14,745,903
RPL7Ac	Fal25	60S ribosomal protein L7a	21881	17	7,532,190
RPL11	Fal29	Ribosomal protein L11	06305	23	5,837,007
RPL23	Fal17	60S ribosomal protein L23	02496	27	3,957,106
RPL30c	Fal1	60S ribosomal protein L30	36621	2	132,392,442
RPL37Ac	Fal7	60S ribosomal protein L37a	18702	7	24,873,612
SBDS	Fal13	Shwachman–Bodian–Diamond syndrome	01658	19	782,450
SERINC1	Fal2	Tumor differentially expressed 2	23989	3	63,939,040
SPCS1	Fal21	Signal peptidase complex subunit 1	02526	12	749,082
STYX	Unlinked	Serine/threonine/tyrosine-interacting protein	20280	5	60,813,050
TGFB2c	Fal2	Transforming growth factor β-2 precursor	15664	3	20,477,096
THUMPD3	Fal31	THUMP domain-containing protein 3	13674	12	20,067,036
TIMM17A	Fal27	Translocase of inner mitochondrial membrane 17 homolog A	00123	26	1,096,567
TMc	Fal23	Tropomyosin α-1 chain	05572	10	5,107,950
TMEM32	Fal18	Transmembrane protein 32 precursor	09949	4	4,186,140
TUBGCP3	Fal4	Gamma-tubulin complex component 3	27189	1	142,174,289
TXNDC14	Unlinked	Thioredoxin domain containing 14	11887	5	18,159,850
UBE2J1	Fal2	Ubiquitin-conjugating enzyme E2, J1	25442	3	78,396,237
UCHL3	Fal4	Ubiquitin carboxyl-terminal esterase L3	27356	1	159,089,890
UCHL5	Fal11	Ubiquitin carboxyl-terminal hydrolase L5	03977	8	3,533,310
UQCRC1	Fal21	Ubiquinol-cytochrome-c reductase complex core protein 1	09300	12	9,301,215
VIPR2	Fal1	Vasoactive intestinal peptide receptor 2	10623	2	9,569,988
VISL1	Unlinked	Visinin-like protein 1	26565	3	102,957,139
VPS26A	Unlinked	Vacuolar protein sorting-associated protein 26A	06635	6	11,919,165
WDR24	Fal28	WD repeat protein 24	03862	14	13,908,481
YME1L1	Fal1	YME1-like 1	12112	2	15,827,680
YPEL5	Unlinked	Yippee-like 5	14726	3	8,031,083
Microsatellites
EST9	Fal10			6	24,700,943
EST10	Fal10			6	24,725,801
EST16	Fal1	Adenylate cyclase-activating polypeptide 1		2	104,980,406
EST31	Fal19	Peroxiredoxin-6		8	4,336,642
EST46	Fal30	α-Fetoprotein enhancer-binding protein		11	21,546,200
Fhy215	Fal1
Fhy216	Fal3
Fhy217	Fal3
Fhy220	Fal4
Fhy221	Fal16
Fhy223	Fal4	NW_001471545.1 Gga1 WGA43_2		1	140,906,192
Fhy224	Fal9	NW_001471459.1 Gga15 WGA207_2		15	1,723,096
Fhy225	Fal2
Fhy226	Fal2
Fhy227	Fal8
Fhy228	Fal22
Fhy230	Fal1	NW_001471633.1 Gga2 WGA60_2		2	49,531,823
Fhy231	Unlinked
Fhy234	Fal1
Fhy235	Fal22	NW_001471676.1 Gga3 WGA95_2		3	5,902,333
Fhy236	Fal15	NW_001471568.1 Gga20 WGA258_2		20	12,551,523
Fhy237	Fal6
Fhy301	Fal16	NW_001471434.1 Gga11 WGA182_2		11	4,504,384
Fhy304	Fal5
Fhy306	Fal5
Fhy310	Fal17
Fhy321	Fal10
Fhy326	Fal6	NW_001471698.1 Gga5 WGA124_2		5	16,318,616
Fhy328	Fal5
Fhy329	Fal2	NW_001471669.1 Gga3 WGA102_2		3	48,212,536
Fhy336	Fal1
Fhy339	Fal8
Fhy341	Fal3
Fhy342	Unlinked
Fhy344	Fal5	NW_001471687.1 Gga4 WGA113_2		4	84,661,433
Fhy350	Fal9	NW_001471461.1 Gga15 WGA209_2		15	9,498,662
Fhy356	Fal3	NW_001471552.1 Gga1 WGA4_2		1	7,812,491
Fhy361	Fal1	NW_001471633 Gga2 WGA60_2		2	45,075,848
Fhy370	Fal8	NW_001471529.1 Gga1 WGA29_2		1	92,668,432
Fhy401	Fal5
Fhy403	Fal1
Fhy404	Unlinked
Fhy405	Fal9
Fhy407	Fal4	NW_001471554.1 Gga1 WGA51_2		1	177,252,431
Fhy408	Fal3
Fhy413	Fal1
Fhy415	Fal1	NW_001471639.1 Gga2 WGA66_2		2	79,942,990
Fhy427	Fal3	NW_001471510.1 Gga1 WGA11_2		1	21,414,296
Fhy428	Fal1
Fhy429	Unlinked
Fhy431	Fal10
Fhy432	Unlinked
Fhy444	Fal4
Fhy448	Fal2
Fhy450	Fal7
Fhy452	Fal4
Fhy453	Fal10	NW_001471713.1 Gga6 WGA139_2		6	3,534,436
Fhy454	Unlinked
Fhy458	Fal1	NW_001471651.1 Gga2 WGA78_2		2	129,424,241
Fhy464	Fal2
Fhy465	Fal4	NW_001471554.1 Gga1 WGA51_2		1	176,719,296
Fhy466	Fal7	NW_001471729.1 Gga7 WGA155_2		7	13,437,385
Fhy467	Fal1
FhU3	Fal1
FhU4	Fal3
FhU5	Fal1
GG-C25	Fal6			5	31,720,224
SS12	Fal3
ZF-C59	Fal1	Adenylate cyclase-activating polypeptide 1		2	104,980,320
ZF-S8	Unlinked	Hypothetical protein		18	4,041,446
ZF-S9	Fal12			13	8,811,918

a

The contig number from the chicken genome sequence build 2.1 is described for microsatellites.

b

ENSEMBL ID for the orthologous chicken gene (ENSGALG000000xxxxx).

c

Loci from Borge  et al. (2005).

Seventy microsatellites were isolated from the closely related pied flycatcher (Ficedula hypoleuca: Leder  et al. 2008). Sixty-three of these markers, as well as five EST-linked microsatellites and nine microsatellites from other passerines (Karaiskou  et al. 2008), were PCR multiplexed in sets of six to nine loci using 30 ng of DNA per reaction. Each PCR multiplex could be analyzed on a single run of an ABI3130xl (Applied Biosystems). Detailed PCR multiplex protocols and electrophoresis details can be found in Karaiskou and Primmer (2007).

Data analysis:

Microsatellites were scored using the GeneMapper software (Applied Biosystems). SNPs from the same intron were combined into haplotypes using the available pedigree information. For both types of markers, missing data points of parents were inferred from the haplotypes of offspring and mates when possible. Offspring showing deviations from the expected parental genotypes/haplotypes were not included in further analysis. These are likely to represent extra-pair offspring since previous work in this population has revealed that ∼15% of all offspring result from extra-pair copulations (Sheldon  et al. 1997; Sheldon and Ellegren 1999).

Linkage analyses were performed with CRI-MAP (Green  et al. 1990). Initially, all markers were tested against each other with the two-point option and markers that clustered together with significant lod score support (>3.0) were treated as linkage groups. Framework maps were constructed with the build option and the best position of all markers within an ordered linkage group was then estimated with recurrent runs of the option flips4 until no better order could be found (best order map).

Microsatellite clone sequences were used in cross-species MEGABLAST searches against the chicken genome sequence (http://www.ncbi.nlm.nih.gov/genome/seq/BlastGen/BlastGen.cgi?taxid=9031), first with default settings and then with relaxed settings according to Dawson  et al. (2006). Both search methods generated the same set of significant (arbitrarily set at <E⁻⁵) hits, although the relaxed settings generated somewhat lower E-values. Because of this procedure 20 of the microsatellites could be anchored to a single location in the chicken genome and therefore be used for comparative studies. Chicken genomic locations were taken from Build 2.1 of the chicken genome, obtained from http://www.ncbi.nlm.nih.gov/mapview/maps.cgi?taxid=9031. Data on chicken recombination rates were from the same source. The particular markers that we used for collared flycatcher mapping were not always included in the chicken genetic map. To obtain recombination rate estimates for orthologous regions of the collared flycatcher and chicken genomes, we therefore used data from genetic markers located preferably within 1 Mb of our markers in the chicken genome. Graphical presentations of genetic maps were created in MapChart (Voorrips 2002).

RESULTS

Marker analysis:

We genotyped 384 previously identified SNPs in a pedigree of 322 collared flycatchers. After excluding SNPs found to be monomorphic in the pedigree (36 sites) or with a sample call rate <60% (23), segregation data for 321 SNPs from 170 genes became available. These genes are from 26 of the 28 chromosomes contained within the most recent (May 2006) version of the chicken genome assembly; it was not possible to develop conserved markers for Gga16 and Gga25 due to a lack of assigned genes to these small microchromosomes. SNPs from the same intron were combined into haplotypes to increase the informativeness of each gene in pedigree analysis. In addition, 71 polymorphic microsatellites were scored in the pedigree. We thereby had a total of 241 loci (“markers”) available for mapping. The average number of informative meioses for SNP haplotypes and microsatellites was 184 and 240, respectively (supplemental Table 2).

Linkage mapping:

Linkage analysis detected significant two-point linkage (lod score >3.0) to at least one other marker for 211 of the 241 loci. These 211 markers form a best-order genetic map consisting of 33 linkage groups (Figure 1), ranging in size from a genetic length of 325 cM (with 37 markers) to 0 cM (two groups with two markers showing no recombination). We refer to these linkage groups as Fal1-33 (F. albicollis linkage groups 1–33). The mean genetic distance between adjacent markers in the map is 10.0 cM (±6.7 SD). A framework map based on markers showing a multi-point lod score >3 is presented in supplemental Figure 1; generally, the framework map differs from the best-order map only by the inability to confidently place some markers on either side of adjacent loci.

Figure 1.—

A genetic linkage map of the collared flycatcher genome. Cumulative genetic distances in centimorgans are given to the right of each chromosome. Gene markers (genotyped by SNPs) are in regular text while anonymous microsatellites are in italics.

The total sex-average length of the best-order map is 1787 cM, with significantly more recombination in males (P = 0.01, Wilcoxon's test for paired data). Overall, recombination is 22% higher in males than in females (total distance 1982 vs. 1627 cM), although for individual intervals the female point estimate is sometimes higher than the male estimate. Sex-specific linkage maps are presented in supplemental Figure 2.

Thirty markers (22 genes and eight microsatellites) remain unlinked. It is clear that this is at least in part due to a lack of power to detect linkage, since unlinked markers are highly overrepresented among markers with a low number of informative meioses (supplemental Table 2). However, nine unlinked genic SNP haplotypes and six unlinked microsatellites appear as singletons despite >70 informative meioses. Because the microsatellites were isolated by random library screening (Leder  et al. 2008), it is possible that they correspond to regions of the chicken genome, which is not yet contained within the chicken genome assembly. For the gene-based SNPs, a possible explanation of the absence of linkage is that they are located in regions of high recombination within conserved linkage groups. Given the observation that recombination rates tend to be elevated toward telomeric ends of avian chromosomes (International  Chicken  Genome  Sequencing  Consortium 2004; Schmid  et al. 2005; Wahlberg  et al. 2007), this explanation is supported by the fact that 6 of 9 unlinked genes are the most distal marker on chicken chromosomes in our marker set (at 2p, 3p, 4q, 5q, 10p, and 10q, respectively). Alternatively, they may indicate chromosomal rearrangements (see supplemental data).

Comparative mapping:

A major advantage of this map is that genic markers allow the identification of homologous regions in the genomes of other species; this is particularly important for the ability to transfer genomic information from model organisms to nonmodel ones. The 33 collared flycatcher linkage groups correspond to 24 different chromosomes in chicken, with the 2 remaining chicken chromosomes from which we had markers (Gga18 and Gga22) each being represented by two to three unlinked singletons. By comparing the location of gene sequences in the chicken and collared flycatcher genomes (as well as of 20 flycatcher microsatellites for which a homologous sequence could be identified in chicken), we find a remarkable degree of chromosomal conservation, both at the level of shared synteny and at the level of conserved gene order (Figure 2). No fewer than 18 chicken chromosomes correspond to a single linkage group in the collared flycatcher, indicating completely conserved synteny.

Figure 2.—

Genome conservation revealed by the alignment of orthologous collared flycatcher linkage groups and chicken chromosomes. Chromosomes/linkage groups are drawn proportional to their genetic length (the same for both species) with the full length of chicken chromosomes displayed.

Seven chicken chromosomes are orthologous to two to three collared flycatcher linkage groups, suggestive of fusion/fission events. However, as discussed in detail in the supplemental data, several of these cases are likely to represent a lack of power to connect linkage groups that originate from the same collared flycatcher chromosome. The interval between SNP-based gene markers is generally <10 Mb in the chicken genome. As recombination rates are known to vary across the genome, particularly due to the presence of recombination hot spots, physical intervals of this size may correspond to genetic distances longer than that possible for detecting linkage. Of course, length expansions in the collared flycatcher genome would accentuate such problems. In the end, only two rearrangements that distinguish the karyotypes of the chicken and the collared flycatcher are strongly supported by our data (supplemental data) and are also independently confirmed in other species. These include a fission of the ancestral chromosome 1 in Passeriformes to yield the collared flycatcher linkage groups Fal3 and Fal8 (supported by data in Guttenbach  et al. 2003; Derjusheva  et al. 2004; Itoh and Arnold 2005) and the fusion of ancestral chromosomes 4 (corresponding to Fal5) and 10 (Fal18) in the chicken lineage to yield Gga4 (supported by Reed  et al. 2005). One or a few more rearrangements could be indicated by our data but would need a denser map for confirmation.

Generally, gene order within syntenic groups is completely conserved in the chicken–collared flycatcher comparison. There are 18 cases of intrachromosomal rearrangements in the best-order map. However, only 7 of these remain in the framework map [corresponding to Gga1-Fal4 (two cases), Gga1-Fal8, Gga3-Fal2, Gga4-Fal5, Gga15-Fal9, and Gga20-Fal15; supplemental Figure 1]; the other 11 cases are in the form of inversed order of markers that is not statistically supported (multi-point lod score <3). Six of the 7 well-supported cases are consistent with rather short inversions of ∼1.5–6 Mb according to the chicken genome sequence, while the seventh represents a large proportion of Gga4-Fal5 that spans at least 38 Mb.

DISCUSSION

This work reveals an evolutionary stasis of gene order and chromosome organization in a comparison of two highly diverged avian lineages, estimated to have separated 100 million years ago (MYA) ago. Our results confirm the stability of avian genomes previously indicated by results from comparative mapping in chicken and mammals (Burt  et al. 1999; Bourque  et al. 2005) and from cytogenetic (e.g., Griffin  et al. 2007) and genetic (e.g., Dawson  et al. 2007) analyses within birds. However, as previous work has been largely insensitive to intrachromosomal structures, and to fine-scale interchromosomal organization, our findings extend the observation of the genomic stability of birds to be valid also at the level of gene order and organization and across distantly related taxa.

Stasis of avian genome organization:

The karyotype of the collared flycatcher has not been characterized, but the great majority of passerine birds so far analyzed have a chromosome number between 78 and 80, i.e., quite similar to the 2n = 78 of chicken (Gregory  et al. 2006). The structure of the karyotype is also well conserved with a few large macrochromosomes and a large number of very small microchromosomes. Moreover, available data from measurements of DNA content indicate that genome size is more or less identical in chicken and birds from the order Passeriformes, at just ≈1 Gb (Gregory  et al. 2006). Assembled genome sequence has been assigned to 28 of the 36 chicken autosomes (Gga29-36 are not covered in the assembly and represent microchromosomes with <5–10 Mb DNA). Most of these chromosomes correspond to a single linkage group in the collared flycatcher map. We find strong support for two interchromosomal rearrangements that have occurred since the divergence of the lineages leading to the chicken and the collared flycatcher, rearrangements which are also supported by previous observations (International  Chicken  Genome  Sequencing  Consortium 2004; Griffin  et al. 2007). There are other gaps in the collared flycatcher linkage map aligned to chicken that might indicate the presence of one to three additional fissions or fusions (corresponding to Gga1, Gga3, and Gga12). However, it is possible that we lack the power to detect linkage in these cases because markers flanking the gaps might be too far apart, contain too little information, or are separated by a region of high recombination rate (see supplemental data).

There is also evidence for at least seven intrachromosomal rearrangements in the chicken–collared flycatcher comparison. Four of these have occurred in the vicinity of the sites for interchromosomal rearrangement, indicating that there are regions of avian chromosomes that are particularly fragile and prone to different types of rearrangement. The number of observed intrachromosomal rearrangements in the collared flycatcher–chicken comparison is roughly two to three times the number of interchromosomal rearrangements. A similar ratio is also found in mammals (Pevzner and Tesler 2003; Pontius  et al. 2007). This indicates that it is not just the inter- or the intrachromosomal rate of rearrangement that is low in birds; rather, it is the overall rate of rearrangements that is uniformly low.

The incidence of inter- and intrachromosomal rearrangements seen in this study can be used to roughly estimate the minimum rate of chromosomal rearrangement in birds. We have to take into account that our map does not cover the entire genome and that the resolution may be insufficient for detecting small rearrangements within mapped regions. With this caveat in mind and assuming a 66% genomic coverage of the map (see below) and 100 MY since the most recent common ancestor of Galliformes and Passeriformes (200 MY of evolution in total), there is an overall minimum rate of chromosomal exchange during avian evolution of approximately one event/15 (200/9 × 0.66) MY (0.045/MY). Separate estimates for inter- and intrachromosomal rearrangement give one fission/fusion event/66 MY (0.015/MY) and one large-scale inversion event/19 MY (0.05/MY). This is considerably lower than what has been estimated for mammalian lineages (≈0.6–2.0; Pontius  et al. 2007) and gives a quantitative idea of the relative stability of avian genomes.

The genomic integrity shown by the comparison of two distantly related avian lineages provides insight into vertebrate karyotype evolution. It has been shown that the rate of chromosomal diversification varies considerably on a temporal scale (Ferguson-Smith and Trifonov 2007). For example, while the rate of chromosomal diversification is thought to have been comparatively high in the lineage leading from an early vertebrate ancestor to the eutherian ancestor, by contrast it was considered to be low in the lineage from the eutherian ancestor to the primate ancestor (and continued to be so in the human lineage) (Kohn  et al. 2006). However, no such temporal heterogeneity is indicated in either the chicken or the collared flycatcher lineage. In fact, the ancestral vertebrate karyotype is highly conserved in the chicken so genomic stability must have been prevalent all along the lineage leading to modern birds since the split of synapsids and diapsids 310 MYA. The only clear exception is provided by birds of prey that have an atypical bird karyotype that lacks distinct macrochromosomes and microchromosomes (Deoliveira  et al. 2005).

Why does the degree of genomic integrity differ extensively between birds and mammals? One possible explanation relates to the role of repeat elements in governing chromosome fragility and the dearth of active families of interspersed repeat elements in avian genomes (International  Chicken  Genome  Sequencing  Consortium 2004). Breakpoints for chromosome rearrangements tend to be enriched with interspersed repetitive elements, low-copy repeats, and segmental duplications (Bailey  et al. 2004; Freudenreich 2007; Kehrer-Sawatzki and Cooper 2007), including in chicken (Gordon  et al. 2007). Mechanistically, unequal crossing over between repeated sequences within or between chromosomes is likely to contribute to this pattern (Lupski and Stankiewicz 2005). The bird genome has remained compact and repeat poor during avian evolution (Organ  et al. 2007); there is only one abundant class of interspersed repeats in the chicken genome: the long interspersed CR1 element, with a vanishingly small proportion representing anything other than short and heavily truncated copies (International  Chicken  Genome  Sequencing  Consortium 2004). In light of the potential link between a low-repeat content and chromosomal integrity, it is tempting to invoke a nonadaptive explanation to genomic stability of birds.

One interesting aspect of genomic stability of birds is the role of chromosomal rearrangements in speciation (Noor  et al. 2001; Rieseberg 2001; Navarro and Barton 2003). It is known that crosses between chromosomal variants can result in hybrid inviability or impaired fitness (Capanna and Castiglia 2004). Speciation rates differ considerably between birds and mammals with the evolution of post-zygotic incompatibility being ∼10 times slower in birds than in mammals (Price and Bouvier 2002; Fitzpatrick 2004). It is possible that the formation of post-zygotic barriers in birds is delayed due to the slow rate of chromosomal rearrangement.

A partial genetic map, based mainly on microsatellites and AFLP markers, has been developed for another passerine bird, the great reed warbler (Hansson  et al. 2005; Åkesson  et al. 2007; Dawson  et al. 2007). The seven collard flycatcher–chicken gene-order differences that we detected in this study are not evident in the comparison between the great reed warbler and the chicken (Dawson  et al. 2007). This could suggest that these inversions arose in the collared flycatcher lineage subsequent to the split of Muscicapoidea–Sylvoidea. However, the physical marker density and chromosomal coverage in the great reed warbler map is relatively low, especially for markers informative in comparative mapping, so it may be that the resolution is not sufficient for detection of some intrachromosomal rearrangements. Neither of these inversions is seen in the fairly dense map of the turkey genome (Reed  et al. 2005) so they may have arisen either early in the evolution of Galliformes or somewhere along the passerine lineage leading to the collared flycatcher. The great reed warbler map revealed two clear cases of inversions when compared to chicken chromosomes 1 and 2, respectively (Dawson  et al. 2007). However, they are not seen in the collared flycatcher–chicken comparison despite dense marker coverage in these regions. They are therefore likely to have arisen in the lineage leading to the great reed warbler subsequent to the split of Muscicapoidea–Sylvoidea.

Sex-specific recombination rates:

There is evidence for heterochiasmy in collared flycatchers and the direction (22% more recombination in males) is in agreement with the Haldane–Huxley rule, which states that, in species where the autosomal recombination rate differs quantitatively between sexes, it is usually the heterogametic sex that has a reduced recombination rate. However, as in several galliform birds (Groenen  et al. 2000; Kayang  et al. 2004; Reed  et al. 2005), the difference is not as pronounced as that seen in most mammals [in humans, for example, the female rate is ∼1.6 times higher than the male rate (Broman  et al. 1998)]. The great reed warbler shows a quite contrasting pattern, with a more than twofold excess of recombination in females (Hansson  et al. 2005; Dawson  et al. 2007), i.e., against the expectations of the Haldane–Huxley rule. This difference between two passerine birds is unexpected and not easily conceived in light of alternative hypotheses for the evolution of heterochiasmy (Lenormand 2003). For example, there is some support for the idea that haploid selection plays a role (Lenormand and Dutheil 2005). According to this model, and assuming epistasis, the sex experiencing the largest variance in reproductive success should recombine less to keep favorable gene combinations together. However, both collared flycatchers (Sheldon  et al. 1997; Sheldon and Ellegren 1999) and great reed warblers (Hasselquist  et al. 1998) are polygynous species and there is no obvious reason to believe that the intensity of gametic selection would differ between them. As it stands now, further work in additional bird species is needed to obtain a more broad-scale picture of avian recombination rates.

Genome coverage:

How large a proportion of the collared flycatcher genome is covered by this map? One way to approach this question is to make use of physical information from the chicken genome. The amount of chicken sequence contained within linkage groups shared between the chicken and the collared flycatcher is 663 Mb, or 66% of the total ≈1-Gb autosomal genome (the sequence assigned to chromosomes in the chicken genome assembly covers ∼90% of the genome). Obviously, additional sequence is covered in the flycatcher map if the 15 highly polymorphic unlinked singletons are included and when considering the fact that each end marker in all linkage groups covers some flanking distance. Assuming a similar genome size as in chicken (Gregory  et al. 2006), it may be estimated that our map covers 75–80% of the collared flycatcher genome. It is more difficult to assess the proportion of the total genetic length of the flycatcher genome covered by the map. First, we probably lack markers from a number of microchromosomes to which assembled sequence has not yet been assigned in the chicken. However small a microchromosome is, it contributes 50 cM to the total map length since there is an obligate crossing over per chromosome (Jones and Franklin 2006). Second, recombination rates tend to increase toward telomeres, so it can therefore be difficult to assess the amount of recombination outside linkage groups by extrapolation. When it comes to genetic coverage, we thus conclude that it must be less than the physical coverage estimated above.

Genomics of natural populations:

There is a slow but steady progress in applying genetic maps for quantitative trait locus (QTL) mapping in natural animal populations (Slate 2005), either directly in the wild or by using wild-caught individuals raised in the laboratory. Notable examples from ecologically important animal models include the analysis of the genetic basis of body armour in the stickleback Gasterosteus aculeatus (Peichel  et al. 2001; Colosimo  et al. 2005) and of body weight and morphology in the free-living Soay sheep Ovis aries (Beraldi  et al. 2007) and the red deer Cervus elaphus (Slate  et al. 2002). In the two latter systems, microsatellite markers developed in closely related domestic animals could be used across species. In sticklebacks, anonymous, species-specific microsatellites were used and random approaches for map development have also included the use of microsatellites, RAPD, or AFLP markers in a number of other species (Reid  et al. 2007; Troggio  et al. 2007).The development of the collared flycatcher map represents an application of a new approach to genetic mapping in natural populations. For several reasons, it benefited from an earlier systematic effort to develop a set of conserved reference markers evenly spread across the avian genome (Backström  et al. 2008).

First, through identification of particularly conserved sites for primer design, these markers had a high amplification success rate in collared flycatchers. Second, the design of intronic amplicons 500–1000 bp in size uncovered >900 polymorphic sites with just a moderate amount of screening (10 unrelated individuals; Backström  et al. 2008), from which we could choose the most informative polymorphisms for genotyping. SNP-based mapping is particularly attractive since large-scale analysis can be conducted using automatic procedures; this study generated 150,000 genotypes. Third, and importantly, since the reference set of markers was developed to represent a uniform coverage of all assembled chicken chromosomes, the markers also covered a significant part of the collared flycatcher genome. Clearly, using the same number of randomly selected markers would not have given the same degree of genome coverage. Indeed, the high degree of synteny and gene-order conservation in birds added to this benefit, although, in principle, the use of markers evenly distributed in a related species should always give better coverage than using random markers. Fourth, as the first map of a natural population based on protein-coding gene markers, it is straightforward to transfer genetic information from model species by means of anchoring of orthologous loci. These features should apply equally to any bird species and we therefore envision that the marker set will become important for future studies of the genetics of natural bird populations. The use of the same set of markers should also aid comparison of maps across many different species. More generally, the approach of designing evenly distributed, conserved gene markers from one or more model species could potentially be applied to many groups of organisms.

While SNPs offer a nearly inexhaustible source of genetic markers that are amenable to large-scale analysis, they come with the price of typically showing lower levels of polymorphism compared to microsatellites. For this reason, we also included a set of microsatellites in this study, as this was expected to facilitate anchoring of linkage groups. However, the difference in average information content between microsatellites and SNPs was not very large (mean number of informative meioses of 240 and 184, respectively). Moreover, we could not detect any significant difference in the percentage of unlinked markers from the different categories (8 of 71 for microsatellites vs. 22 of 170 for SNP haplotypes). This is likely due to the fact that for most genes we combined SNPs into haplotypes, thereby increasing their informativeness.

To conclude, we report here a gene-based genetic map of the collared flycatcher that, when compared to the chicken map, demonstrates extensive synteny and gene-order conservation during 100 MY of evolution. This high degree of genome stability of birds is likely to have important consequences for their general evolutionary patterns, including processes of speciation.

Footnotes

1

Present address: Department of Genetics, Development and Molecular Biology, School of Biology, Aristotle University of Thessaloniki, P.O. Box 54, 124 Thessaloniki, Macedonia, Greece.

Footnotes

Communicating editor: N. A. Jenkins

Acknowledgement

Financial support was obtained from the Swedish Research Council. SNP genotyping was performed at the SNP Technology Platform at Uppsala University, which is gratefully acknowledged.