Chunk #36 — Methods — Generalized linear models

Source

Statistical modeling for sensitive detection of low-frequency single nucleotide variants.

Embedded

yes

Text

The zero-inflated negative binomial distribution can be written as:5\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ P\left({N}_{s,b,l}={n}_{s,b,l}\Big|{\boldsymbol{c}}_{s,b,l},\ {\boldsymbol{z}}_{s,b,l}\right)=\Big\{\begin{array}{c}\hfill {\pi}_{s,b,l}+\left(1-{\pi}_{s,b,l}\right)NB\left({\mu}_{s,b,l},\ \theta; 0\right)\kern2.75em if\kern0.5em {n}_{s,b,l}=0\hfill \\ {}\hfill \left(1-{\pi}_{s,b,l}\right)NB\left({\mu}_{s,b,l},\theta; {n}_{s,b,l}\right)\kern4.75em if\kern0.5em {n}_{s,b,l}>0\hfill \end{array} $$\end{document}PNs,b,l=ns,b,l|cs,b,l,zs,b,l={πs,b,l+1−πs,b,lNBμs,b,l,θ;0ifns,b,l=01−πs,b,lNBμs,b,lθns,b,lifns,b,l>0