Incorrect outputs of population imputation for very small sample input

[XingerTang]

Same inputs as #51, with population only imputation command:

AlphaImpute2 -genotypes tests/functional_tests/test_2/genotypes.txt -pedigree tests/functional_tests/test_2/pedigree.txt -pop_only -phase_output -seg_output -out tests/functional_tests/outputs/test_2

Both the genotypes output:

1 0 0 0 0 0
2 0 0 0 0 0
3 0 0 0 0 0
4 0 0 0 0 0

and

1 0 0 0 0 0
1 0 0 0 0 0
2 0 0 0 0 0
2 0 0 0 0 0
3 0 0 0 0 0
3 0 0 0 0 0
4 0 0 0 0 0
4 0 0 0 0 0

are all zeros, while the outputs of the combined imputation command:

AlphaImpute2 -genotypes tests/functional_tests/test_2/genotypes.txt -pedigree tests/functional_tests/test_2/pedigree.txt -phase_output -seg_output -out tests/functional_tests/outputs/test_2

are

1 0 0 0 0 0
2 0 0 0 0 0
3 1 0 0 0 0
4 0 0 0 0 0

and

1 0 0 0 0 0
1 0 0 0 0 0
2 0 0 0 0 0
2 0 0 0 0 0
3 0 0 0 0 0
3 1 0 0 0 0
4 0 0 0 0 0
4 0 0 0 0 0

Has only one locus corrected.

[XingerTang]

In 4ce7465, error messages:

Fewer than 10 HD individuals found for population imputation. Population imputation cannot proceed. Consider using -ped_only imputation or reducing -hd_threshold parameter.

are added to warn the user when there are fewer than 10 HD individuals, as in that case, the population imputation would perform poorly.

[gregorgorjanc]

This is a useful error message! How did you come up with the number 10? I don't want you to get stuck on this number and funding an optimum value, but we need to communicate that even 10 is likely too small. Please change the error message to this effect. Thanks!

[XingerTang]

@gregorgorjanc
The number 10 is chosen arbitrarily. I have modified the error message to not emphasise the number 10:

Too few HD individuals were found for population imputation. Population imputation cannot proceed. Consider using -ped_only imputation or reducing the -hd_threshold parameter.

The check that triggers this error message is still 10 HD individuals.

[gregorgorjanc]

Please change to

Too few HD individuals were found for population imputation (the filter is set at 10 HD individuals, but you likely need way more!). Population imputation cannot proceed. Consider using -ped_only imputation or reducing the -hd_threshold parameter.

[XingerTang]

@gregorgorjanc Thank you! The error message is updated.

[gregorgorjanc]

@XingerTang I assume this solves this issue or is the error coming from elsewhere?!

[XingerTang]

@gregorgorjanc Yes, this would solve the issue. Too few HD individuals would affect the calculation of the population imputation and lead to unreliable results. Adding this error message would prevent the user from obtaining the results that are incorrect or unexpected.

[gregorgorjanc]

Grand! Add a note to ChangeLog, close and merge. We should move to you opening PRs and I review PRs instead of commits on your end;)