Description Usage Arguments Details Value Author(s) See Also Examples

The function does an ANOVA of cell survival data from experimental 2-way designs where a treatment factor is tested on a control and on an altered cell line or where two different simultaneous treatments are tested on cells from a common unaltered clone. The function is a wrapper for the R-function `glm`

. quasipoisson family is used with link function `"log"`

, i.e. dependency of treatment factors is considered as logarithmically additive.

1 | ```
cfa2way(X, A, B, param="A/B", method="ml")
``` |

`X` |
a data frame which contains columns |

`A` |
a character string containing the name of a treatment or cell line variable (first factor in the model) |

`B` |
a character string containing the name of a treatment or cell line variable (second factor in the model) |

`param` |
Controls the parametrization of the model. Options are "A/B" for B nested in A , "B/A" for A nested in B and "A*B" for interaction term. |

`method` |
determines the method used for the fit. |

In the data frame `X`

, `Exp`

identifies the experimental replicates and may be numeric or non-numeric. The two treatment or cell line columns should have numeric values 0, 1, ... for 2, 3, ... levels. For e.g. if a column describes clonal alteration (transfection, knock-down etc.) by a gene then 0 means unaltered or control and 1 means altered. Similar if a column describes treatment with one dose then 0 means untreated and 1 treated. 2 would indicate another dose level from the same treatment drug without taking it as a continuous covariate as for cell survival curves for radiation.

The function returns an object of class `cfa2way`

containing three elements, `fit1`

, `fit2`

and `anv`

. `fit1`

and `fit2`

are objects of class `glm`

when `method="ml"`

or of class `lm`

when `method="ls"`

. `fit1`

has logarithmic additive parameters without interaction. `fit2`

has logarithmic additive parameters and interaction. `anv`

is of class `anova`

and contains the F-test. The full result is returned invisibly, i.e. the function has to be used with `print`

or assigned to a variable, say for e.g. `fitcomp`

as in the example below.

Herbert Braselmann

`glm`

and `family`

with references for generalized linear modelling.

1 2 3 4 5 | ```
datatab<- read.table(system.file("doc", "exp2_2waycfa.txt", package="CFAssay"), header=TRUE, sep="\t")
names(datatab) # has columns "x5fuCis" and "siRNA"
fitcomp<- cfa2way(datatab, A="siRNA", B="x5fuCis", param="A/B")
print(fitcomp, labels=c(A="siRNA",B="x5fuCis"))
print(cfa2way(datatab, A="siRNA", B="x5fuCis", param="A/B", method="ls"))
``` |

```
[1] "Exp" "x5fuCis" "siRNA" "ncells" "ncolonies"
*** Two-way ANOVA for factors A and B with interaction ***
A= siRNA , B= x5fuCis
Test for interaction: F-test
F Pr(>F)
values 9.831 0.01202 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Use 'print' to see detailed results
*** Logarithmic linear two-way ANOVA for factors A and B with interaction ***
=============================================================================
A= siRNA , B= x5fuCis
Postscript digits for A or B: 0 inactive, 1 active
surv_percent = exp(Estimate)*100
Null hypothesis (Model 1): no interaction
-----------------------------------------
Estimate Std. Error t value Pr(>|t|) surv_percent
A1 -0.4237844 0.06366051 -6.656944 5.660071e-05 65.5
B1 -1.1187559 0.07738539 -14.456939 4.980588e-08 32.7
Goodness-of-fit values
Residual Deviance: 77.99569
Total sum of squared weighted residuals ssqwresTot: 75.87275
Residual Degrees of Freedom: 10
Dispersion parameter: 7.587275
Alternative hypothesis (Model 2): interaction
---------------------------------------------
parametrization: A/B
Estimate Std. Error t value Pr(>|t|) surv_percent
A1 -0.3484218 0.05266374 -6.615972 9.746888e-05 70.6
A0:B1 -0.9757699 0.07170380 -13.608343 2.619891e-07 37.7
A1:B1 -1.3432352 0.09470322 -14.183627 1.832336e-07 26.1
Goodness-of-fit values
Residual Deviance: 37.15596
Total sum of squared weighted residuals ssqwresTot: 37.38767
Residual Degrees of Freedom: 9
Dispersion parameter: 4.154185
Analysis of Variance Table and F-test
Model 2 versus Model 1
Resid. Df Resid. Dev Df Deviance F Pr(>F)
1 10 77.996
2 9 37.156 1 40.84 9.831 0.01202 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
*** Two-way ANOVA for factors A and B with interaction ***
A= siRNA , B= x5fuCis
Test for interaction: F-test
F Pr(>F)
values 8.1026 0.0192 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Use 'print' to see detailed results
*** Logarithmic linear two-way ANOVA for factors A and B with interaction ***
=============================================================================
A= A , B= B
Postscript digits for A or B: 0 inactive, 1 active
surv_percent = exp(Estimate)*100
Null hypothesis (Model 1): no interaction
-----------------------------------------
Estimate Std. Error t value Pr(>|t|) surv_percent
A1 -0.4952308 0.08463741 -5.851205 1.613459e-04 60.9
B1 -1.1591090 0.08463741 -13.694997 8.357008e-08 31.4
Goodness-of-fit values
Total sum of squared residuals ssqresTot: 0.2865396
Residual Degrees of Freedom: 10
Multiple R-squared: 0.9925945
Alternative hypothesis (Model 2): interaction
---------------------------------------------
parametrization: A/B
Estimate Std. Error t value Pr(>|t|) surv_percent
A1 -0.3110078 0.09152632 -3.398015 7.898129e-03 73.3
A0:B1 -0.9748860 0.09152632 -10.651428 2.111054e-06 37.7
A1:B1 -1.3433320 0.09152632 -14.677001 1.362680e-07 26.1
Goodness-of-fit values
Total sum of squared residuals ssqresTot: 0.1507872
Residual Degrees of Freedom: 9
Multiple R-squared: 0.996103
Analysis of Variance Table and F-test
Model 2 versus Model 1
Res.Df RSS Df Sum of Sq F Pr(>F)
1 10 0.28654
2 9 0.15079 1 0.13575 8.1026 0.0192 *
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
```

Embedding an R snippet on your website

Add the following code to your website.

For more information on customizing the embed code, read Embedding Snippets.