2025-02-02
tamidata2:PRJNA218851
data.frame
en donde incluimos la variable Stage
y el conteo correspondiente a este gen. count Stage
SRR975551Aligned.out.sam.bam 539 Cancer
SRR975552Aligned.out.sam.bam 563 Cancer
SRR975553Aligned.out.sam.bam 1018 Cancer
SRR975554Aligned.out.sam.bam 393 Cancer
SRR975555Aligned.out.sam.bam 398 Cancer
SRR975556Aligned.out.sam.bam 672 Cancer
Call:
glm(formula = count ~ Stage, family = poisson(link = log), data = df)
Coefficients:
Estimate Std. Error z value Pr(>|z|)
(Intercept) 6.729957 0.008146 826.14 <2e-16 ***
StageMetastasis -0.306800 0.012512 -24.52 <2e-16 ***
StageNormal 0.429249 0.010467 41.01 <2e-16 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for poisson family taken to be 1)
Null deviance: 13236.9 on 53 degrees of freedom
Residual deviance: 8723.7 on 51 degrees of freedom
AIC: 9189.2
Number of Fisher Scoring iterations: 4
La función de probabilidad es \[ f(y; \phi, \mu) = \frac{\Gamma(y+\phi)}{\Gamma(\phi)\Gamma(y+1)} \bigg ( \frac{\phi}{\mu+\phi} \bigg)^\phi \bigg ( 1- \frac{\phi}{\mu+\phi} \bigg)^y \] con \(y= 0, 1, 2 , \ldots\) donde \(\phi\) y \(\mu\) son los parámetros.
Se tiene que \[ E(Y) = \mu, \] \[ var(Y) = \mu + \frac{\mu^2}{\phi}. \]
El parámetro \(1/\phi\) es un parámetro de dispersión.
Call:
MASS::glm.nb(formula = count ~ Stage, data = df, init.theta = 5.191786539,
link = log)
Coefficients:
Estimate Std. Error z value Pr(>|z|)
(Intercept) 6.7300 0.1038 64.858 < 2e-16 ***
StageMetastasis -0.3068 0.1468 -2.090 0.03666 *
StageNormal 0.4292 0.1467 2.927 0.00343 **
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
(Dispersion parameter for Negative Binomial(5.1918) family taken to be 1)
Null deviance: 81.344 on 53 degrees of freedom
Residual deviance: 55.733 on 51 degrees of freedom
AIC: 796.61
Number of Fisher Scoring iterations: 1
Theta: 5.192
Std. Err.: 0.976
2 x log-likelihood: -788.611