Mailman 3 - fastlmm-user

New bed-reader feature: Support for writing very large PLINK 1.9 Bed Files
by Carl K 22 Apr '24

22 Apr '24

Greetings, The Python version of the bed-reader package (1.0.3) now supports writing out large PLINK 1.9 Bed Files too big to fit in memory. Documentation on the new features: https://fastlmm.github.io/bed-reader/#bed_reader.create_bed Here is a notebook that shows the creation of a 1M sample x 250K SNP Bed file (In just, 5 minutes). Also, you can convert between individual-major to SNP-major mode (In about 1 hour). https://github.com/fastlmm/bed-reader/blob/master/bed_reader/tests/benchmar… Thanks to Fabian G. who suggested this feature<https://github.com/fastlmm/bed-reader/issues/25>. Yours, Carl Carl Kadie, Ph.D. FaST-LMM & PySnpTools Team<https://fastlmm.github.io/> (Microsoft Research, retired) https://www.linkedin.com/in/carlk/ Join the FaST-LMM user discussion and announcement list via email<mailto:fastlmm-user-join@python.org?subject=Subscribe> (or use web sign up<https://mail.python.org/mailman3/lists/fastlmm-user.python.org>)

1 0

New version of Bed-Reader and PySnpTools
by Carl K 12 Nov '23

12 Nov '23

Greetings, We're happy to announce new versions of Bed-Reader (1.0.0) and PySnpTools (0.5.11): Bed-reader The Python version of Bed-Reader now supports SciPy's sparse matrix format. Also, by default, it now depends only on NumPy. If you want that minimal version use pip install bed-reader If you want support for downloading example files and creating sparse matrices use: pip install bed-reader[samples,sparse] Rust version: The Rust version of Bed-Reader now returns a smaller "result code". (Result<_,Box<BedErrorPlus>> instead of Result<_, BedErrorPlus>.) PySnpTools We cleaned up the code, removing references to Python 2. Also, updated some sample files to fix warnings that they were compressed with Python 2 methods. * Carl Carl Kadie, Ph.D. FaST-LMM & PySnpTools Team<https://fastlmm.github.io/> (Microsoft Research, retired) https://www.linkedin.com/in/carlk/ Join the FaST-LMM user discussion and announcement list via email<mailto:fastlmm-user-join@python.org?subject=Subscribe> (or use web sign up<https://mail.python.org/mailman3/lists/fastlmm-user.python.org>)

1 0

New version: FaST-LMM 0.6.7
by Carl K 11 Aug '23

11 Aug '23

Greetings, We are happy to announce a new version of FaST-LMM (https://pypi.org/project/fastlmm/). It now supports Python 3.11 and newer versions of NumPy and SciPy. * Carl Carl Kadie, Ph.D. FaST-LMM & PySnpTools Team<https://fastlmm.github.io/> (Microsoft Research, retired) https://www.linkedin.com/in/carlk/ Join the FaST-LMM user discussion and announcement list via email<mailto:fastlmm-user-join@python.org?subject=Subscribe> (or use web sign up<https://mail.python.org/mailman3/lists/fastlmm-user.python.org>)

1 0

Bed-Reader and PySnpTools now support Mac on Arm (and also Python 3.11)
by Carl K 21 May '23

21 May '23

Greetings, We're happy to announce that Bed-Reader and PySnpTools now support MacOS Arm when installed via "pip". Also, we've bumped our supported Python versions up so we now support Python 3.8, 3.9, 3.10, and 3.11. Yours, Carl Carl Kadie, Ph.D. FaST-LMM & PySnpTools Team<https://fastlmm.github.io/> (Microsoft Research, retired) https://www.linkedin.com/in/carlk/ Join the FaST-LMM user discussion and announcement list via email<mailto:fastlmm-user-join@python.org?subject=Subscribe> (or use web sign up<https://mail.python.org/mailman3/lists/fastlmm-user.python.org>)

1 0

New versions of FaST-LMM, PySnpTools, and Bed-Reader
by Carl K 02 May '23

02 May '23

Greetings, We're released a new version of the tools. The main change is adding support for Python 3.10 on Windows. * bed-reader 0.2.34<https://pypi.org/project/bed-reader/> - Updated to use latest version of Rust tools that we use under the covers. (Also updated: article about creating Python extensions in Rust<https://medium.com/towards-data-science/nine-rules-for-writing-python-exten…>.) * pysnptools 0.5.9<https://pypi.org/project/pysnptools/>, now supports Python 3.10 on Windows (except that Bgen format can't currently be used on Windows and Python 3.10) * fastlmm 0.6.5<https://pypi.org/project/fastlmm/>, now supports Python 3.10 on Windows This means that all the tools now support Python 3.7, 3.8, 3.9, and 3.10 on Linux, Windows, and MacOS. * Carl Carl Kadie, Ph.D. FaST-LMM & PySnpTools Team<https://fastlmm.github.io/> (Microsoft Research, retired) https://www.linkedin.com/in/carlk/ Join the FaST-LMM user discussion and announcement list via email<mailto:fastlmm-user-join@python.org?subject=Subscribe> (or use web sign up<https://mail.python.org/mailman3/lists/fastlmm-user.python.org>)

1 0

New version of PySnpTools: 0.5.8
by Carl K 01 Feb '23

01 Feb '23

Greetings, We've updated PySnpTools to work with NumPy 1.24. (PySnpTools is also used by FaST-LMM.) pip uninstall pysnptools pip install pysnptools Yours, Carl Carl Kadie, Ph.D. FaST-LMM & PySnpTools Team<https://fastlmm.github.io/> (Microsoft Research, retired) https://www.linkedin.com/in/carlk/ Join the FaST-LMM user discussion and announcement list via email<mailto:fastlmm-user-join@python.org?subject=Subscribe> (or use web sign up<https://nam10.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.pyth…>)

1 0

FaST-LMM adds "show_snp_fract_var_exp" option
by Carl K 31 Oct '22

31 Oct '22

Greetings, FaST-LMM 0.6.4 is out. See fastlmm * PyPI<https://pypi.org/project/fastlmm/> The new version adds show_snp_fract_var_exp as a new option to single_snp. To install: pip uninstall fastlmm pip install fastlmm Details: The output will always show 'EffectSize'. If show_snp_fract_var_exp is True, then the output will also show 'SnpFractVarExp'. EffectSize and SnpFractVarExp are different measures of fraction of variance explained. The first uses all variance of the phenotype in the denominator, whereas the second uses variance of the phenotype after removing the effects of population structure and family relatedness. Specifically, EffectSize is SNPWeight^2 * "raw" test SNP variance / phenotype variance, where: * SNPWeight is the beta for the "standardized test SNP" (where "standardized" means made to have mean 0, stdev 1) * "raw" means the original test SNP values of "0,1,2" (the count of allele 1 or 2 [it doesn't matter]). * the phenotype's original values are used (so, they are not, for example, changed by regressing out the covariates). Yours, Carl Carl Kadie, Ph.D. FaST-LMM & PySnpTools Team<https://fastlmm.github.io/> (Microsoft Research, retired) https://www.linkedin.com/in/carlk/ Join the FaST-LMM user discussion and announcement list via email<mailto:fastlmm-user-join@python.org?subject=Subscribe> (or use web sign up<https://nam10.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.pyth…>)

1 0

FaST-LMM and PySnpTools support for Python 3.10
by Carl K 23 Oct '22

23 Oct '22

Greetings, We're happy to announce support for Python 3.10 for the Linux and MacOS versions of FaST-LMM and PySnpTools (but, sadly, not Windows). The new versions also fixes a problem with reading the BGEN format on newer Python version. To install: pip uninstall fastlmm pip uninstall pysnptools pip install pysnptools pip install fastlmm Yours, Carl Carl Kadie, Ph.D. FaST-LMM & PySnpTools Team<https://fastlmm.github.io/> (Microsoft Research, retired) https://www.linkedin.com/in/carlk/ Join the FaST-LMM user discussion and announcement list via email<mailto:fastlmm-user-join@python.org?subject=Subscribe> (or use web sign up<https://nam10.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.pyth…>)

1 0

New minor versions of FaST-LMM, PySnpTools, and Bed-Reader
by Carl K 08 Sep '22

08 Sep '22

Greetings, I'm happy to announce new Python versions of FaST-LMM, PySnpTools, and Bed-Reader. These new versions keep the packages up-to-date with NumPy, etc. (If you haven't seen an issue, there is no need to upgrade.) We use the Rust language to make these Python packages faster. If you are interested in Rust for scientific software, please see: * [article] Nine Rules for Elegant Rust Library APIs: Practical Lessons from Porting Bed-Reader, a Bioinformatics Library, from Python to Rust<https://towardsdatascience.com/nine-rules-for-elegant-rust-library-apis-9b9…>, Towards Data Science, June 2022. * [video] Nine Rules for Elegant Rust Library APIs<https://www.youtube.com/watch?v=6-8-9ZV-2WQ>, Seattle Rust Meetup Group, September 2022. * [article] Nine Rules for Writing Python Extensions in Rust: Practical Lessons from Upgrading Bed-Reader, a Python Bioinformatics Package<https://towardsdatascience.com/nine-rules-for-writing-python-extensions-in-…>, Towards Data Science, December 2021. * [video] Nine Rules for Writing Python Extensions with Rust & Rayon<https://www.youtube.com/watch?v=oIxYsR47q3Q&t=2355s>, Seattle Rust Meetup Group, January 2022. * Carl Carl Kadie, Ph.D. FaST-LMM & PySnpTools Team<https://fastlmm.github.io/> (Microsoft Research, retired) https://www.linkedin.com/in/carlk/ Join the FaST-LMM user discussion and announcement list via email<mailto:fastlmm-user-join@python.org?subject=Subscribe> (or use web sign up<https://nam10.safelinks.protection.outlook.com/?url=https%3A%2F%2Fmail.pyth…>)

1 0

FaST-LMM version 0.6.1 now with Effect Size
by Carl K 22 Jun '22

22 Jun '22

Greetings, We're pleased to announce the release of FaST-LMM 0.6.1 The single_snp function now returns the effect size of the SNP. EffectSize is SNPWeight^2 * "raw" test SNP variance / phenotype variance, where * SNPWeight is the beta for the "standardized test SNP" (where "standardized" means made to have mean 0, stdev 1) * "raw" means the original test SNP values of "0,1,2" (the count of allele 1 or 2 [it doesn't matter]) * The phenotype's original values are used (so, they are not, for example, changed by regressing out the covariates To install: pip uninstall fastlmm pip install fastlmm * Carl

1 0