Question regarding the video from Riffomonas Project

[adriaaula]

Hi,

I have been reading the literature regarding this topic, and the video from Pat Schloss showed that sequencing depth is still influencing the estimation in breakaway.

https://www.youtube.com/watch?v=xwpMNRt57Zo

How should we proceed in cases like this one? Or what is wrong with his approach?

Thanks a ton!

[adriaaula]

Playing around with the examples he provides I found that through the creation of the null distribution the data structure is skewed, losing most of the zeroes, and modifying the structure of the ASV count distribution, even though it keeps the row sums and the column sums equally.

And therefore also affecting the index by frequency table?

Plot generated through the code Pat kindly shared in his github: https://github.com/riffomonas/distances/tree/0f64c79deabf30603205ffc983abc6674cd121b4

This could explain the weird distribution, but I would love to discuss it!