new method to compute T from C

[imitxelena003]

Zimmerman et al. JCTC 2018, eq. 8

Compute transition probability matrix by a adding a pseudocount to count matrix and normalize it.
This method avoids creating sources or sinks, so it turns out to be adequate for adaptive sampling calculations.

[imitxelena003]

The new method to estimate T is helpful when doing adaptive sampling calculations.

TICA is not finished, but it already has the necessary tools implemented.

[daviddesancho]

Maybe worth finding out whether the reference to Zimmerman, 2018 is the most relevant one? Check, e.g. eq. 55 in Prinz et al, JCP, 2011