Data integration in microbial genomics ... - Jacobs University

More documents

Recommendations

Info

82 6. Domains of unknown function Figure 6.1: Hypothesis generation using network representations of Spearman rank correlations between Domain of Unknown Function (DUF) abundances across GOS metagenomes. Vertices are labelled with the corresponding DUF number (i.e. “59” represents DUF59). Network I includes DUFs with a variety of functions, possibly involved in a response to nutrient input in the marine epipelagic zone. Network II is dominated by DUFs linked to photosynthetic organisms and functions. Network III is composed of three DUFs known to co-occur. a) Fruchterman- Reingold embedded network. Edge lengths are inversely related to correlation strength b) Minimum spanning tree representation of DUF correlations. Only edges describing the shortest path (hence, strongest correlation) between vertices are visualized. EamA (PF00892, formerly ‘DUF6’) domain suggests that its activity may play a role in controlling this functional constellation. Hypotheses generated from covariation across metagenomes may be contextualized with environmental data to enhance interpretation. We
6.4. Results 83 employed indirect gradient analysis (see [Ramette, 2007] for a review of multivariate analyses) after [Virtanen et al., 2006] to relate DUF abundances in network II (Fig. 6.1; 17 DUF varieties) to chlorophyll concentrations at appropriate GOS sites (n=56; see Supplementary methods online). We standardized DUF abundances at each site by the median abundance of 22 ‘single-copy domains’ detected at that site. Figure 6.2: Hypothesis contextualization using indirect gradient analysis of GOS sites described by the abundances of DUFs in Network II (ref. Fig. 6.1) and chlorophyll concentration data. Each bubble represents one GOS site and bubble size reflects the in situ chlorophyll concentration measured during the expedition. Bubble positions reflect the Bray-Curtis dissimilarity between sites calculated from the per site abundance profiles of DUFs in Network II. The blue vector describes the linear fit of chlorophyll concentration data to the ordination (Rv 2 ≈ 0.52, P (> R) ≈ 9.99 × 10 − 4). Red isoclines describe a generalized additive model fit of the same chlorophyll data to the ordination (Rs 2 ≈ 0.91, p ≈ 2.00 × 10 − 16) and are labeled with the corresponding chlorophyll concentration (in 4µg chlorophyll per kg seawater). Regions where the chlorophyll isoclines and vector intersect perpendicularly suggest a coherent response gradient of metagenomic DUF content to chlorophyll concentrations. We then ordinated sites by non-metric dimensional scaling (NMDS; Fig. 6.2, hollow circles) using Bray-Curtis dissimilarities. Next, we
Page 1:
Max Planck Institute for Marine Mic
Page 4 and 5:
List of abbreviations BLAST Basic L
Page 6 and 7:
Thesis abstract Deoxyribonucleic ac
Page 8 and 9:
4.2 Introduction . . . . . . . . .
Page 10 and 11:
2 1. Introduction syntax: data leve
Page 12 and 13:
4 1. Introduction used to derive ne
Page 14 and 15:
6 1. Introduction croorganisms thri
Page 16 and 17:
8 1. Introduction the middle at the
Page 18 and 19:
10 1. Introduction marker genes hel
Page 20 and 21:
12 1. Introduction what kind and wh
Page 22 and 23:
14 1. Introduction possibilities li
Page 24 and 25:
16 1. Introduction tual data acquis
Page 26 and 27:
18 1. Introduction Doug Wendel, Owe
Page 29 and 30:
CHAPTER 2 METABAR A tool for consis
Page 31 and 32:
2.2. Background 23 molecular sequen
Page 33 and 34:
2.2. Background 25 printing of data
Page 35 and 36:
2.4. Results 27 KML export function
Page 37 and 38:
2.4. Results 29 Figure 2.3: Screens
Page 39 and 40: 2.5. Discussion 31 taking samples i
Page 41 and 42: 2.5. Discussion 33 Handlebar MetaBa
Page 43 and 44: 2.6. Conclusion 35 rine Microbiolog
Page 45: 2.6. Conclusion 37 RK advised progr
Page 48 and 49: 40 3. CDinFusion Genomic Standards
Page 50 and 51: 42 3. CDinFusion sortium (GSC), an
Page 52 and 53: 44 3. CDinFusion tiFASTA file with
Page 54 and 55: 46 3. CDinFusion offered the option
Page 56 and 57: 48 3. CDinFusion AMD Opteron TM pro
Page 58 and 59: 50 3. CDinFusion able separately an
Page 60 and 61: 52 3. CDinFusion and upload it with
Page 63 and 64: CHAPTER 4 MIMARKS The Minimum infor
Page 65 and 66: 4.2. Introduction 57 edge generatio
Page 67 and 68: 4.3. Development of MIMARKS 59 in t
Page 69 and 70: 4.4. Survey of published parameters
Page 71 and 72: 4.5. The MIMARKS checklist 63 Resul
Page 73 and 74: 4.6. Conclusions and call for actio
Page 75 and 76: CHAPTER 5 MEGX.NET Integrated datab
Page 77 and 78: 5.3. New database structure and con
Page 79 and 80: 5.3. New database structure and con
Page 81 and 82: 5.4. User Access 73 5.4 User Access
Page 83 and 84: 5.4. User Access 75 tal context, co
Page 85: 5.5. Summary 77 access charge: Max
Page 88 and 89: 80 6. Domains of unknown function 6
Page 92 and 93: 84 6. Domains of unknown function p
Page 95 and 96: CHAPTER 7 SUMMARY AND DISCUSSION Th
Page 97 and 98: 7.2. GSC standards development 89 T
Page 99 and 100: 7.2. GSC standards development 91 o
Page 101 and 102: 7.4. In silico Hypothesis Generatio
Page 103: 7.5. Getting the most out of the da
Page 106 and 107: 98 8. Conclusion and outlook nities
Page 108 and 109: 100 8. Conclusion and outlook plina
Page 110 and 111: 102 BIBLIOGRAPHY [Beynon-Davies, 20
Page 112 and 113: 104 BIBLIOGRAPHY oceanic regions by
Page 114 and 115: 106 BIBLIOGRAPHY [Giovannoni et al.
Page 116 and 117: 108 BIBLIOGRAPHY [Kitano, 2002] Kit
Page 118 and 119: 110 BIBLIOGRAPHY pathways in marine
Page 120 and 121: 112 BIBLIOGRAPHY [Ramette, 2007] Ra
Page 122 and 123: 114 BIBLIOGRAPHY (2006). Microbial
Page 124 and 125: 116 BIBLIOGRAPHY [Williamson et al.
Page 127: CHAPTER 9 ACKNOWLEDGEMENTS First, I
show all

Data integration in microbial genomics ... - Jacobs University

Create successful ePaper yourself

Delete template?

Save as template?