Encounter

Thursday, December 04, 2008

Acceleration of Mitochondrial Free Radical Theory of Aging

Experiments in silico:

22 cycles of
{ 2^[nth cycle] loops of
{SSA loops till W>7000
+ Hypergeometric mtDNA division
}
}

2^22 cycles of
{SSA loops till timecount>980
}

Initial attempt was to replace the SSA loops with Tau-leaping, and SSA loop reduction by generating and using reaction probability based on smaller sample size.

But it turns out that because the reaction isn't as complicated as anticipated with only 2 reaction routes and reaction species, SSA might be faster than the above 'accelerations' after all...

The reaction time 980 may also not be substantially large enough to make tau-leaping cost effective, especially due to tau-leaping's need to generate the poisson random number per loop.

The large number of cells (2^22= 4 194 304) appears to be the largest source of computational time... Need some way to circumvent this.