Update T conversion to generations and years

Author: tjstruck

docs/userguide/demog.md

@@ -6,15 +6,22 @@ One of the primary ways demographic models in dadi are distinguished from one an
 
 Demographic models with more than one population (may be refered to elsewhere in documentation as 2D or 3D for two or three populations respectivly) can have `m`, the migration rate, as a parameter. It is common in dadi documentation to denote migration rate by `m`, then the destination population, followed by the source population. For example, `m12` would translate to the rate of population 2 that migrates to population 1.
 
-<!-- `nu` is the population size relative to a reference population (the ancestral population or the effective population size). So a `nu` of 3 means the population is triple the size of the reference population, or \\(3 N_{\text{ref}}\\).
+`nu` is the population size relative to a reference population (the ancestral population or the effective population size). So a `nu` of 3 means the population is triple the size of the reference population, or \\(3 N_{\text{ref}}\\).
 
 `T` is the time in \\(\text{ploidy} \cdot N_{\text{ref}}\\) generations. So for humans, time is in units of `T` are \\(2 N_{\text{ref}}\\)
-In order to convert `T` into years, for diploids, users would use the conversion: \\(2 N_{\text{ref}} T\\)
-To convert `T` into generations, users would divide years by generations per-year, ex. for Humans estimating 25 years per-generation: \\(\frac{2 N_{\text{ref}} T}{25 \text{ years per generation}}\\).
-
-Each model inference will produce a \\(\theta\\) value, which is roughly a population scale neutral mutation rate. This value is important for estimating the DFE (see [dadi documentation](https://dadi.readthedocs.io/en/latest/user-guide/dfe-inference/) for more specific details) and calculating \\(N_\text{ref}\\), using the conversion \\(\frac{\theta}{4 \mu L}\\), where \\(\mu\\) is the genomic mutation rate and \\(L\\) is the length of sequence that could have ended up in the SNPs data. Put another way, \\(L\\) is the total length of the genome that was sequenced and could have been the same type of SNP (intergenic, synonymous, nonsynonymous) being analyized. -->
-
-<!-- For demographic models with more than one population, `m`, the rate of migration, is an additional parameter that is common in dadi demographic models. It is common in dadi documentation to denote migration rate by `m`, then the destination population, followed by source population. For example, `m12` would translate to the rate of population 2 that migrants to population 1. `m` can be converted into units of fraction of inviduals in a destination population made up of a source population: \\(\frac{m}{2 N_{\text{ref}}}\\) -->
+In order to convert `T` into generations, for diploids, users would use the conversion: \\(2 N_{\text{ref}} T\\)
+To convert `T` into years, users would multiply by years per-generation, 
+ex. for Humans estimating 25 years per-generation: \\(2 N_{\text{ref}} T \times \frac{25 \text{years}{\text{generation}}}\\).
+
+Each model inference will produce a \\(\theta\\) value, which is roughly a population scale neutral mutation rate. 
+This value is important for estimating the DFE (see [dadi documentation](https://dadi.readthedocs.io/en/latest/user-guide/dfe-inference/)
+for more specific details) and calculating \\(N_\text{ref}\\), 
+using the conversion \\(\frac{\theta}{4 \mu L}\\), where \\(\mu\\) is the genomic mutation rate 
+and \\(L\\) is the length of sequence that could have ended up in the SNPs data. 
+Put another way, \\(L\\) is the total length of the genome that was sequenced and could have been the same type of SNP 
+(intergenic, synonymous, nonsynonymous) being analyized.
+
+For demographic models with more than one population, `m`, the rate of migration, is an additional parameter that is common in dadi demographic models. It is common in dadi documentation to denote migration rate by `m`, then the destination population, followed by source population. For example, `m12` would translate to the rate of population 2 that migrants to population 1. `m` can be converted into units of fraction of inviduals in a destination population made up of a source population: \\(\frac{m}{2 N_{\text{ref}}}\\) -->
 
 Another potential parameter users might infer is `misid`, the percentage of ancestral misidentification. If users' SNP data contains ancestral allele state (usually in the VCF, this will be denoted as AA= in the INFO column), dadi-cli can generate an unfolded allele frequency spectrum by assuming the derived allele is the one not matching the ancestral state rather than assuming the derived allele is the one with the lower population frequency. This results in more SNPs that are shared in more of the population. `misid` corrects for model assumptions that SNPs with high population prevalence are rarer.