Reader Comments

Post a new comment on this article

On division time asymmetry estimation

Posted by julou on 23 Mar 2011 at 14:00 GMT

Hello,

In my experience, estimating when division happens using transmitted light picture can be highly subjective. In addition, depending on acquisition frequency, there is always an error on division time…

As you do not comment very much on how you procede in this study, I would be interested in knowing more about your method. This appears critical to me as all your analysis is based on this division time asymmetry.

Thank you in advance for your attention. Best regards,

Thomas Julou

No competing interests declared.

RE: On division time asymmetry estimation

mwatve replied to julou on 28 Mar 2011 at 09:18 GMT

Thanks for your interest. In the first phase of the study we were indeed worried by the subjectivity and possible observer bias in recording division time. During an earlier phase of the study reported in Nature Precedings earlier (1) we performed double blind trials to estimate observer variation and possible bias in recording division time. To estimate observer errors a series of photomicrographs were taken on 10 developing microcolonies each with an interval of 1 minute and the series was viewed and the time of division independently recorded by five different observers. In this phase we were using transmitted light microscopy. There was a maximum difference of 3 minutes in recording division time and the error distribution was approximately normal. The error was still considerably smaller than the range of differences in division times of daughter cells which had a maximum over 50 minutes. The differences between selection and substrate concentrations were significant after double blinding the trial design. Also blinded and open experiments did not very significantly.
In the second phase of the study reported in this paper we shifted to DIC (differential interference contrast) microscopy which increased the contrast and clarity of the images so much that individual variation came down to less than 1 minute (please go through the set of images provided in the supplementary material). Also during the second phase for a given experiment all the readings were taken by the same person so as to avoid individual difference. Admittedly, the 1 min resolution limits the precision of recording. However distribution of observed differences between division times is such that 1 min resolution is sufficient. Most of the statistical significance between groups is contributed by division time differences between 3-30 minutes. Therefore, although your concern is theoretically correct it hardly makes any difference for the statistical inferences.

1. Watve, Milind, Watve, Samit, Belsare, Prajakta, Lele, Uttara, Bari, Snehal, and Karande, Sarika. Caloric restriction causes symmetric cell division and delays aging in Escherichia coli. Available from Nature Precedings <http://hdl.handle.net/101...> (2008)

No competing interests declared.