The genetic architecture of adaptive radiation in Heliconius melpomene

Lead Research Organisation: University of Cambridge
Department Name: Zoology


Abstracts are not currently available in GtR for all funded research. This is normally because the abstract was not required at the time of proposal submission, but may be because it included sensitive information such as personal details.

Technical Summary

Heliconius erato and H. melpomene have evolved over 20 distinct geographic races that are convergent between the species due to mimicry, offering an excellent opportunity to study the genetic basis of phenotypic adaptation. Here I will identify the genomic regions that underlie the radiation of H. melpomene using a high-density genetic linkage map. Candidate genes, chosen as likely to be involved in wing pattern and pigment formation or identified as being differentially expressed between different pigment types, will be included in the map. Homologous linkage groups will be identified in the mimetic species, Heliconius erato, to determine whether the same genomic regions are used to produce convergent colour patterns in mimetic species.


10 25 50
Description Objective 1 Sequence ESTs from developing wing tissue. EST data sets are now available for H. melpomene and H. erato largely derived from wing tissue. In total we have obtained 17573 sequences from a single H. erato combined-stage wing disc library and 4976 for H. melpomene derived from pupal wing disc libraries. These are clustered, annotated and publicly available from our web database at Some difficulties in generating high quality libraries from wing cDNA from H. melpomene led us to concentrate sequencing efforts at the beginning of the grant on the H. erato library, and these were undertaken by our collaborators in the US. In the meantime however we were able to proceed with the linkage mapping aspects of the project by taking advantage of the H. erato sequences. Subsequently however we have managed to prepare high quality normalised cDNA pools from wing tissue which have been sequenced using novel 454 technology, partly funded by this grant. Thus, we now have sequences for perhaps around 9000 H. melpomene genes expressed during wing development (far more than could possibly have been envisaged at the start of this project). These are not yet submitted to a public database as they only became available over the last month of the project and have not yet been analysed fully, however as with all data generated under this project they will be made publicly available once processed and will greatly increase public resources for lepidopteran genomics. Thus by the end of the project we have greatly exceeded the expected amount of cDNA sequence obtained for gene finding in Heliconius melpomene.

Importantly, this work led us to establish a Lepidoptera-wide EST database ( that now forms a resource for international work on Lepidoptera genomics. This relational database makes available all lepidopteran EST data in a clustered and annotated format. Currently the database houses 273,077 mRNA sequences, clustered into 70,867 gene objects most of which also have predicted protein sequences. This contrasts with only 6,907 protein sequences held on GenBank for all of the Lepidoptera, many of which are partial and fragmented sequences. All of the sequences have been processed using a common analysis pipeline [6], making the analysis directly comparable between species. These data are searchable both using text searches against databased similarity annotation, and by BLAST. Several BLAST tools are available including the original blastall options (BLASTn, tBLASTx etc) and newer algorithms psi-BLAST and msBLAST. The latter is designed for searching using short protein sequences obtained from proteomics studies. This database was hosted by us at the University of Edinburgh until late 2007, and its long term security after termination of this grant was ensured by moving it to the Max Planck Institute in Jena where it is currently hosted by the laboratory of David Heckel. Further development of the database has also moved to Jena primarily driven by Alexie Papanicolaou who was a Masters student with the PI in Edinburgh before moving to Jena.

Objective 2 Develop high density linkage maps and map patterning loci. We have published a high-density linkage map for H. melpomene that spans the 21 linkage groups over 1616 cM and contains 261 markers [2]. These include AFLP, microsatellite and cDNA derived markers. Since publication of that paper we have developed many more markers from our EST sequences [3]. To date, over 70 single copy nuclear genes and 22 microsatellite markers have been mapped. We did not quite achieve the 200 mapped markers as originally proposed, but the results were of considerable significance, demonstrating for the first time a high degree of synteny conservation between H. melpomene and the distantly related moth, Bombyx mori (published in Genetics, [5]). Given the phylogenetic distance between these species, this pattern is likely to hold true across the macrolepidoptera. Thus, the linkage maps generated as part of this project will provide an important resource for gene finding across the whole Lepidoptera.

We have also targeted the development of markers to colour pattern loci. This initially focussed on the Yb locus and has led to the development of the BAC contig described in the proposal. More recently, we have been concentrating on developing markers linked to the B/D loci that also have a major phenotypic effect by controlling the red elements in H. melpomene crosses. So far nine AFLP markers have been identified in this region, and several of these have been sequenced leading to assembly of a BAC tile pathe spanning around 600kb (Figure 1).

In addition to mapping Mendelian factors of major effect we have also been investigating quantitative variation that is segregating in our crosses. We have identified a major effect QTL on linkage group 13 that affects the width of the red forewing band in crosses between H. melpomene cythera and H. melpomene melpomene (Figure 2). This work is currently in review in Heredity.

Objective 3 Identify differentially expressed genes. The third aim outlined in the proposal was to carry out subtractive hybridisation between different races of H. melpomene. We have not so far carried out this experiment, but instead have staged wing development and carried out rt-PCR experiments to study the expression of candidate ommochrome pathway genes (Vermilion, Cinnabar and A-tet-like) in different wing regions. This has demonstrated that the gene Cinnabar is up-regulated specifically in the forewing band of H. melpomene cythera, a pattern that is paralleled in the co-mimic species H. erato. This represents the first evidence of parallel pattern-specific gene regulation in mimetic butterflies (Roberts et al. in prep.).

Objective 4 Map candidate loci. A large number of candidate genes have been identified from our EST sequences based on putative homology with Drosophila or other species. Many of these have now been mapped in our broods, including ommochrome pathway genes (vermilion, scarlet, white, cinnabar), melanin pathway genes (dopa-decarboxylase, yellow-c, yellow-a), signalling molecules (wingless, decapentaplegic) and transcription factors (cubitus interruptus, apterous, engrailed, invected, distal-less, scalloped). None of these candidate loci maps close to any of our colour pattern switch loci [4]. The failure so far of this candidate gene approach is in some ways exciting, as it suggests that novel or at least unexpected loci are likely to be uncovered by the complementary and more direct approach of positional cloning. This work is the subject of a current BBSRC grant.

Objective 5 Compare location of patterning genes between species by means of comparative linkage mapping. Comparative mapping has generated perhaps the most exciting results from this grant so far [4]. We have demonstrated that the Yb locus in H. melpomene maps to the same genomic location as the Cr locus in H. erato and the P locus in H. numata. Furthermore we have shown that the B and D loci in H. melpomene are tightly linked to the gene MRSP, as is the D locus in H. erato that has similar phenotypic effects. This therefore provides evidence that a second locus is homologous between the two species. Thus, by comparing three species, rather than the two originally proposed we have demonstrated conservation of colour pattern regulation that is far greater than we would have envisaged at the start of the project. In particular, the patterns of H. numata are very different from those of H. melpomene and H. erato and it is therefore very surprising that the genetic control in the three species is shared. In addition, H. numata has a 'supergene' controlling polymorphic pattern variation, a genetic system that lay at the centre of the controversy between the 'mutationists' and 'gradualists' in the early part of the 20th century. In this project we have laid the groundwork for elucidating the genetic mechanism of a mimicry 'supergene' for the first time, which will be of considerable interest to all evolutionary biologists.

In summary therefore we have achieved more than anticipated on all of the five major objectives of this grant proposal.
Exploitation Route Further research in this area
Sectors Education,Environment

Description This grant has led to important insights in evolutionary biology that are widely cited in textbooks We have used the results of this work in our presentation for the Royal Society Summer Science Exhibition 2014 in which we described the evolution of butterly wing patterns to ~15000 members of the public
First Year Of Impact 2004
Sector Education,Environment
Impact Types Societal

Title ButterflyBase 
Description EST database 
Type Of Material Database/Collection of data 
Provided To Others? Yes  
Impact Widely used for many years in genomics research