search

MKLau / dissertation

Evolution of Ecological Networks. This is Matt Lau's dissertation projects folder. For more info, please contact him.
1 stars 1 forks source link

Do co-occurrence patterns within a tree change with genotype? #123

Closed MKLau closed 10 years ago

MKLau commented 10 years ago

Assuming co-occurrence influences interactions, then this is a metric for how genotype influences interactions.

SES is not affected by genotype, but multivariate analysis of the separate negative and positive co-occurrence patterns does show a genotypic affect.

MKLau commented 10 years ago

Summary: Senescent leaves show positive co-occurrence patterns, while live leaves show negative co-occurrence patterns.

Positive Co-occurrences Df SumsOfSqs MeanSqs F.Model R2 Pr(>F) pit.g[pit.lf == "live"] 7 2.5657 0.36653 0.99301 0.19888 0.4843 Residuals 28 10.3350 0.36911 0.80112 Total 35 12.9007 1.00000

Df SumsOfSqs MeanSqs F.Model R2 Pr(>F) pit.g[pit.lf == "sen"] 7 3.3563 0.47947 1.4944 0.27199 0.02899 Residuals 28 8.9834 0.32084 0.72801 Total 35 12.3397 1.00000

Negative Co-occurrences Df SumsOfSqs MeanSqs F.Model R2 Pr(>F) pit.g[pit.lf == "live"] 7 1.3405 0.19150 1.8018 0.31056 0.0009998 Residuals 28 2.9759 0.10628 0.68944 Total 35 4.3164 1.00000

Df SumsOfSqs MeanSqs F.Model R2 Pr(>F) pit.g[pit.lf == "sen"] 7 1.7955 0.25650 1.2548 0.23879 0.1346 Residuals 28 5.7237 0.20442 0.76121 Total 35 7.5192 1.00000

MKLau commented 10 years ago

Genotype does not affect SES

cgREML(pit.ses[pit.lf=='live'],pit.g[pit.lf=='live']) chi2 P.value 1.2079907 0.2717304

cgREML(pit.ses[pit.lf=='sen'],pit.g[pit.lf=='sen']) chi2 P.value 2.3845097 0.1225433

cgREML((pit.ses[pit.lf=='sen']-pit.ses[pit.lf=='live']),pit.g[pit.lf=='sen']) chi2 P.value 2.4295155 0.1190697