More Related Content
What's hot
What's hot (14)
Similar to Status seminar 31.01.17, Aalborg University
Similar to Status seminar 31.01.17, Aalborg University (20)
Recently uploaded
Recently uploaded (20)
Status seminar 31.01.17, Aalborg University
- 1. Development of the Next Generation of Bioinformatic Tools in Microbiology Julie K. T. Pedersen Supervisor: Mads Albertsen CENTER FOR MICROBIAL COMMUNITIES 2017-01-31 Aalborg, Denmark
- 2. Agenda CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY • Background • me • data • Aim • Current status • Network • Timeseries • What’s next?
- 3. Who am I? CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY • Back from maternity leave 1. sep • Earlier projects: proteins and DNA • Albertsen Lab (Part of EB group) • Master: purely bioinformatic • Blog: ”Life as a master student”
- 4. Who am I? CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY • Back from maternity leave 1. sep • Earlier projects: proteins and DNA • Albertsen Lab (Part of EB group) • Master: purely bioinformatic • Blog: ”Life as a master student”
- 5. [ ]Human health Salmonella MRSA Y. Pestis Obesity Dental health Immune system
- 7. [ ]Biotechnology Wastewater Biogas Bioplastic Medicaments Drinking water Enzymes
- 8. They help clean our water CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY Wastewater
- 9. Some of them create problems CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY Wastewater
- 10. Identification – 16S rRNA amplicon sequencing CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY Wastewater
- 11. Identification – 16S rRNA amplicon sequencing CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY Wastewater OTU: Operational Taxonomical Unit [Karst et al., 2016]
- 12. Identification – 16S rRNA amplicon sequencing CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY Wastewater [Karst et al., 2016] Name %
- 13. Repeat over time period CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY Wastewater [Karst et al., 2016]
- 14. Time-series from Aalborg West CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY • 9 years, varying sampling frequency • Problems with settling • Filamentous bacteria
- 15. Time-series from AquaDjurs (Fornæs) CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY • Approx. 2 times a week for half a year • Implementing new online control
- 16. Aim CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY Explore microbial correlation patterns in time-series data • Aalborg West • Are there bacteria that correlate with known ”bad” bacteria, either positively or negatively? • Can we identify early warning indicator organisms? • AquaDjurs • How does the bacterial “interactome” look? • How large fraction of the bacteria interact? Writing user-friendly, interactive functions in R • How to make simplicity from complexity?
- 17. CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY
- 18. Learning new skills with DataCamp! CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY
- 19. Learning new skills with DataCamp! CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY
- 20. Learning new skills with DataCamp! CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY
- 21. [ ] Using my new skills! CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY
- 22. Using skills - demo CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY
- 23. CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY What is correlation? Are these 2 bacteria correlated?
- 24. What is correlation? CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY r = 0.816
- 25. What is correlation? CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY R = 0.816 r = 0.816 r = 0.816 r = 0.816 r = 0.816 [Anscombe, 1973]
- 26. What is correlation? CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY Correlation Causation≠
- 27. Looking at correlations CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY Step 1: Calculate a correlation matrix Step 2: Filter and build a network graph Step 3: Make it pretty and identify correlations Step 4: Look at the data behind the correlation
- 28. Step 1: Calculate a correlation matrix CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY A B C D
- 29. Step 1: Calculate a correlation matrix CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY
- 30. Step 2: Build and filter a network graph CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY A B C D 0.82 0.87 0.37 0.96 0.12 0.43
- 31. CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY C 0.87 0.96 A D 0.82 Step 2: Build and filter a network graph
- 32. Step 2: Build and filter a network graph CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY • Thresholds: • OTU Count • Correlation
- 33. Step 3: Make it pretty and identify correlations CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY ”OTU_1” ”OTU_2” Strong correlation between ”OTU_1” and ”OTU_2”?
- 34. Step 4: Look at the data behind the correlation CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY Yes, they are correlated.. r = 0.91
- 35. So, what’s next? CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY http://www.readwritethink.org/files/resources/interactives/timeline_2/
- 36. So, what’s next? CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY • Real data (YEESS!) • Can I find the troublemakers? • Do the troublemakers have followers? • Can i see any indicators using time delay • Other ways to calculate correlation • Can I find tools Open Source (R)? • Will they suggest different correlations? • Which to choose and why?
- 37. Follow my progress CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY Open source: https://github.com/julieklessner www.albertsenlab.org
- 38. Thank you! CENTER FOR MICROBIAL COMMUNITIES | AALBORG UNIVERSITY
Editor's Notes
- Husk at nævne supervisor, snak højere (!)
- Current status on tools and how i got there
- Before leave i did mainly protein and DNA, now I am in the albertsen lab where i do … Purely bioinformatic no lab only sit by my computer very different from snapping funny baby pics. Not in lab data is my starting point where does the data come from? Starting from the top … Bacteria are everywhere
- Purely bioinformatic no lab only sit by my computer very different from snapping funny baby pics. Not in lab data is my starting point where does the data come from? Starting from the top … Bacteria are everywhere
- Often we hear about them from disease, but they are also important for our health
- Mælkesyregæring = fermentation? Lactic acid bacteria Also from food…
- And as I guess you know: Biotechnology
- De hjælper os med at rense vores spildevand. Her er de i et komplekst system. Biological treatment, activated sludgde, specialized qualities.
- Nogle af dem skaber også problemer, men siden det er et komplekst system kan det være svært at vide hvem. Henvis til Peters forklaring af problemer
- Nogle af dem skaber også problemer, men siden det er et komplekst system kan det være svært at vide hvem. Henvis til Peters forklaring af problemer
- Fingerprinting used amplicon sequencing
- Resultaterne af amplicon .. Many more than three
- Repeat over time
- Hvordan påvirkes the microbial community når man anvender de forskellige kontrol pakker.
- Skal være mere direkte linket til vores cases. Har prøvet at gøre det mere “spørgsmålsagtigt” . Du må gerne sætte flere /andre på. Husk at forklare early warning indikator bacteria Evt: joke med at hvis Peter suitcase skal være nyttig skal vi kunne analysere data
- Overgangsslide
- Learning from basics
- Interactive learning (kan måske forklares uden et helt slide)
- To more complex
- Using my skills to combine several lines of code into one. Building functions Detailed analysis Users without in-depth know how Interactivity
- (Kun hvis der er tid) Understreg at det er MIN funktion, gør det personligt Ikke kalde det MiDAS subset, folk udenfor EB kender det ikke.
- Overgangsslide
- Overgangslide
- Forklar at man holder de forskellige tidsserier om imod hinanden og udregner en korrelationskoeffiecient Google “r correlation plot” http://www.gettinggeneticsdone.com/2011/07/scatterplot-matrices-in-r.html
- Understreg at det kun er et udsnit.
- Forklar at hver cirkel er en OTU og linjen representere forholdet i mellem dem.
- Forklar at hver cirkel er en OTU og linjen representere forholdet i mellem dem.
- Må godt være forvirrende: det understreget pointen.
- Forklar at hver cirkel er en bakterie og stregerne imellem representere forholdet imellem dem. + at der i princippet er en linje mellem dem alle, men de er filtreret fra fordi deres forholdet ikke er stærkt nok
- Brug mere tid på at forklare hvad vi ser her. r: Pearsons korrelation koefficient
- Indicators using time delay: correlates when time series is shifted?
- Husk at nævne at bloggen først er startet her januar
- Thanks to the EB group and thanks for listening