deletions | additions       diff --git a/Nuclear area predicts cell cycle phases.tex b/Nuclear area predicts cell cycle phases.tex  index 5f026f3..8dfe27c 100644  --- a/Nuclear area predicts cell cycle phases.tex  +++ b/Nuclear area predicts cell cycle phases.tex 
...
\subsection{Nuclear area predicts cell cycle phases}  In order to validate the use of the nuclear area to infer the cell cycle phases we imaged NIH/3T3 fibroblasts containing the Fucci system. The Fucci system is a two color colors  fluorescent reporter that allow to dinstinguish distinguish  G1 phase from S/G2/M phase \cite{Sakaue-Sawano2008}. We modified our HMM to infer the cell cycle phase from the Fucci signal (denoted $\phi_F$) by multiplying the emmission emission  probabilities of both colors (Figure \ref{fig:cellCycleFucci}A, Section \ref{sec:methods_fucci}) and independently infered inferred  the cell cycle phase $\phi_N$ from the nuclear area (Figure \ref{fig:cellCycleFucci}B). When comparing the two phases we found a reasonable agreement (Figure \ref{fig:cellCycleFucci}C). The two methods are very accurate close to divisions (the abrupt change in the signal at division contains a lot of information) but show more variations in the middle of the cycle.  To provide a more intuitive measure of the precision of our inferrences inferences  we estimated the durations of G1 and S/G2/M from both phases (Figure \ref{fig:cellCycleFucci}D, A-B bottom), overall the S/G2/M (G1) durations infered inferred  from the Fucci signal were predicted from nuclear area with an error of $\pm 2.0h$ ($2.3h$), compared to a mean duration of $14.4h$ ($10.3h$).