Setting the bias-factor

[Ben Cossins]

Hello all,
I have a quick question about the setting of the bias-factor in well-
tempered meta-dynamics. In general the size of barrier to be
encountered is not known before the calculation is started. Hence,
when a barrier that is larger than expected is encountered is it best
to change the bias-factor so that this barrier can be regularly
scaled?
Will this run then still be okay or should another calculation with
a more appropriate bias-factor be started?

Many thanks

ben

[Giovanni Bussi]

Hi Ben,

you are right, the proper bias factor is not know a priori. But also
notice that final convergence is not expected to depend *a lot* on the
bias factor (usually if you are off the right value by 50% everything
will be fine anyway). Also, it is feasible to change it on the fly
after a restart. Just do a normal restart, and plumed will
automatically adjust the previously added hills to the new bias
factor.

Regards,

Giovanni

> --
> You received this message because you are subscribed to the Google Groups "PLUMED users" group.
> To post to this group, send email to plumed...@googlegroups.com.
> To unsubscribe from this group, send email to plumed-users...@googlegroups.com.
> For more options, visit this group at http://groups.google.com/group/plumed-users?hl=en.
>
>

[Ben Cossins]

Hi,
Thanks for your reply Giovanni very helpful. I have one more
question.

In the manual it says that a bias-factor of 10-15 is for systems
with barriers of 6-9 kcal/mol. Is the relationship between the best
bias-factor and the barriers seen in the motion of interest a linear
one? ie. can i choose a bias-factor by scaling the numbers stated in
the manual by the barriers I see in my calculations?

Many thanks for any insight.

Ben

On Nov 11, 7:58 am, Giovanni Bussi <giovanni.bu...@gmail.com> wrote:
> Hi Ben,
>
> you are right, the proper bias factor is not know a priori. But also
> notice that final convergence is not expected to depend *a lot* on the
> bias factor (usually if you are off the right value by 50% everything
> will be fine anyway). Also, it is feasible to change it on the fly
> after a restart. Just do a normal restart, and plumed will
> automatically adjust the previously added hills to the new bias
> factor.
>
> Regards,
>
> Giovanni
>

[Giovanni Bussi]

Yes you can.

Basically, a bias factor of X will reduce your barriers by a factor of
X. Thus, a barrier of 9 kcal/mol, with a bias factor of 15, will
become a barrier of 0.6 kcal/mol, which is easily passed at room
temperature. This is the rationale.

Giovanni