1. Shiri Freilich Janet Thornton’s group, EBI Cambridge University Relating the evolution of gene content to tissue specialization

2. “...And when he was 93, Grandpa decided that the time had come for a man-to-man conversation. (I was 36 at the time, I have been married for 15 years). ‘All my life I am all the time looking at woman, looking and learning. Nu, and what I learned, I want to learn to you now also.’ ‘Woman, in some ways is just like us exactly the same. But in some other ways a woman is entirely different.’ ‘But you know what? In which ways a woman is just like us and in which ways she is very different – nu, on this I am still working’.” (Amos Oz, A Tale of Love and Darkness)

3. The full genome sequence from various species can highlight the common versus unique

4. What’s new?

5. The transition from unicellularity to multicellularity as an interior design challenge - gene

6. Overview Part 1: Expression pattern of ‘old’ and ‘new’ proteins in mouse tissues Part 2: Expression pattern of singleton and duplicate proteins in mouse tissues Part 3: The evolution of tissue-specific metabolic pathways in mammals

7. Part 1: Relating age and function of a protein to its expression pattern in mouse tissues A collaboration with Tom Freeman’s group (MRC RFCGR) Freilich et al, Genome Biol. 2005;6(7):R56.

8. Processing mouse expression and sequence data

9. Classification of mouse proteins into phylogenetic groups

10. Tissues have a similar composition of functional classes Tissue type Fraction enzymes transcription regul. signal transduction transporters

11. Tissues have a similar composition of phylogenetic classes Tissue type Fraction universal eukaryote sp. metazoan sp. mammalian sp.

12. Expression distribution of proteins from different categories Tissues have almost identical functional/ phylogenetic composition Tissue diversity must be achieved through differences in the protein composition within each category Do proteins from different categories duffer in their expression pattern? Number of tissues where protein is expressed Fraction

13. Regulatory proteins are more specifically expressed Number of tissues where is expressed Fraction ~1/3 ~1/10 Mouse proteins classified into functional groups Regulatory proteins Metabolic proteins enzymes transcription regul. signal transduction transporters

14. Metazoan-specific proteins are more specifically expressed Present in a unicellular ancestor of metazoa Number of tissues where protein is expressed Fraction universal eukaryote sp. metazoan sp. mammalian sp. Mouse proteins classified into phyletic groups Specific to metzoa

15. Functional categories overlap with phylogenetic categories Most of the pre-metazoan proteins are metabolic proteins (transporters and enzymes) Most of the metazoan-specific proteins are regulatory proteins (signal transduction and transcription regulation)

16. Identifying the dominant influence: function or age Obvious differences between ‘old’ and ‘new’ proteins, within the metabolic functional group Yet, less than 1/3 of the pre-metazoa proteins are expressed in all tissues Number of tissues where protein is expressed Fraction

17. Still, some of the pre-metazoa proteins are tissue specific Functions occurring in the unicellular cell become tissue-specific in multicellular species (Ldh example) Universal genes that have been duplicated become specific to a tissue whilst a second copy maintains its original expression pattern (Pgk-2 example)

18. Part 2: Relating duplication events to expression pattern in mouse tissues Freilich et al, Genome Biol. 2006;7. “…duplication events had contributed greatly to the attainment of the complex body organisation in metazoa, where cells having identical genetic material can differentiate … due to the presence of duplicated genes in their genomes” Ohno S. (1970). Evolution by gene duplication.

19. The subfunctionalization model Lynch M & Force A, Genetics. 154 (2000): The division of expression of an ancestor gene between its daughter duplicates promotes the retention of a gene in the genome

20. Microarray expression data provide support to the subfunctionalization model Gu et al: expression divergence between duplicate genes increases with evolutionary time (differentiation modes in yeast). Trends Genet. 2002;18: 609-13. Makova et al: spatial expression divergence between duplicate genes increases with evolutionary time (human tissues). Genome Res. 2003;13:1638-45. Huminiecki and Wolfe: a general trend for increased tissue- specificity of expression as family size increase was observed for mammalian genes. Genome Res. 2004;14:1870-79.

21. Project Overview: The relationship between gene duplication and breadth of expression Protein’s perspective: 1.Does duplication event lead to an increase in tissue specificity? Time of duplication perspective: 2.Does the date of duplication event matters? (i.e., do duplication events occurring in the ancestral unicellular lead to an increase in tissue specificity) Protein-family perspective: 3.does a protein family maintain a non-specific expression pattern? (i.e., is a specific expression of proteins from big families complementary) (Freilich et al, Genome Biol. 2006;7(10):R89)

22. Singleton proteins are more globally expressed Number of tissues in which protein is expressed Fraction Singletons (570) Duplicate proteins (1886) Proteins with many close homologues (417)

23. Negative correlation between expression breadth and number of homologues Groups of proteins, ordered by their number of homologues Mean number of tissues Singletons Duplicate proteins

24. Large variation Number of homologues proteins Number of tissues Correlation -0.20 P-value 1.5e-55 Singletons Duplicate proteins Mean number of expressed tissues Correlation -0.20 P-value 1.5e-55

25. Does the date of duplication event matter? (i.e., do duplication events occurring in the ancestral unicellular lead to an increase in tissue specificity) ? Increase in tissue specificity Global expression

26. Identifying ‘old’ and ‘new’ duplications

27. Only post-multicelullarity duplication events lead to expression specificity

28. The protein-family perspective: Is the specific expression of family members complementary? Proteins from big families tend to be more specifically expressed. Does a protein family maintain a non specific expression pattern? ? Complementary expression Overlapping expression

29. Calculating the cumulative tissue distribution of protein families

30. Cumulative tissue distribution of protein families is not correlated with family size Complementary expression pattern in protein families: While a duplication event leads to a tissue specialisation of one or both copy, the total tissue-distribution of the protein family remains constant. Protein families, ordered by size Average tissue-coverage of protein families Singletons Families with any expression information Families with >=75% expression information

31. The findings support the subfunctionalization model Protein’s perspective: 1.Does duplication event lead to an increase in tissue specificity? Yes Time of duplication perspective: 2.Does the date of duplication event maters? only duplication events that that place in a multicellular species lead to a specific expression -> suggests that expression divergence, following gene duplication, promotes the retention of a gene in the genome Protein-family perspective: 3.Does a protein family maintain a non-specific expression pattern? Yes -> suggests the division of expression between family members

32. Part 3: The evolution of the mammalian metabolic pathways Some of the tissue-specific pathways are specific to mammals. Can we understand how tissue-differentiation of animals’ metabolism reflects their evolution? Freilich et al, BMC evolutionary biology 2008, 8:247. STEROID HORMONE METABOLISM

33. Why studying metabolic networks? Metabolic networks’ structure and composition are well defined Available metabolic databases Genotype is highly related to phenotype

34. The structure of the KEGG database STEROID HORMONE METABOLISM

35. The reactions within a pathway can be absent/present in a species Arabidopsis ThalianaHomo Sapiens

36. Project overview: Identification of pathways absent/present in a species Classification of human pathways according to their phyletic origin Characterization of lineage-specific metabolic pathways

37. Phylogenetic classification of human pathways All pathways in human (metabolic) Universal pathways Eukaryota-specific pathways Metazoan-specific pathways Mammalian-specific pathways

38. Components of the Eukaryotic membrane (sphingolipids, glycosaminoglycan) Phylogenetic classification of human pathways 33 Universal pathways 8 Eukaryota-specific pathways 10 Metazoan-specific pathways 14 Mammalian-spc. pathways Metabolic skeleton: sugars, nucleotides, some amino-acids, energy Tissue specific activities: Neuronal guidance, hormonal activity, digestion Tissue specific activities and intracellular signaling (blood cell recognition)

39. The pathways can be linked to form a network

40. The network structure of the metabolic pathways Universal pathways Eukaryota-spc. pathways Metazoan-spc. pathways Mammalian-spc. pathways

41. The integration of the steroid biosynthesis pathway into the sterol biosynthesis pathway sterol cholesterol steroid hormone bile acid Universal Eukaryota Human

42. From manually selected examples to a computational approach Creating a list of adjacent reactions: 2.7.4.2 -> 4.1.1.33 4.1.1.33->2.5.1.1.. 2.5.1.1->2.5.1.21 2.5.1.21->1.14.99.7

43. Using the adjacency list for a large-scale characterization of the metabolic network

44. What’s new (metabolic pathways)?

45. Summary ‘New’ genes tend to be more tissue-specific, ‘ancient’ genes tend to be globally expressed Despite this trend, many metazoan genes are ubiquitous and many universal proteins are tissue specific ‘New’ duplications of ‘old’ and ‘new’ proteins lead to a more specific expression, and therefore can facilitate the evolution of new, tissue-specific, functions The core of metabolic-pathways, inherited from a uniclellular ancestor, provides a platform for the evolution of mammalian- specific, tissue-specific pathways

46. Thanks Thornton Group Janet Thornton Tim Massingham Eric Blanc Expression data: Tom Freeman Sumit Bhattacharyya