Age | Commit message (Collapse) | Author |
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* To help with demonstrating that the code is producing the expected output,
for now, we return the path to the generated html file that displays the
interactive heatmap.
At this point, it is mostly useful in the development environment. Moving
forward, we might have to actually stream the raw html, or if we can get the
Kaleido library packaged for GNU Guix, stream the images binary data instead.
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* Update the plot layouts and size to display the dendrogram and individual
chromosome heatmaps side by side
* Update the colour scale to begin with the grays rather than absolute black
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* Provide the clustering data to be used for the creation of the clustering
dendrogram in the final clustered heatmap plot.
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* Intergrate the heatmap generation code on the /api/heatmaps/clustered
endpoint.
The endpoint should take a json query of the form:
{"traits_names": [ ... ] }
where the "traits_name" value is a list of the full names of traits.
A sample query to the endpoint could be something like the following:
curl -i -X POST "http://localhost:8080/api/heatmaps/clustered" \
-H "Accept: application/json" \
-H "Content-Type: application/json" \
-d '{
"traits_names": [
"UCLA_BXDBXH_CARTILAGE_V2::ILM103710672",
"UCLA_BXDBXH_CARTILAGE_V2::ILM2260338",
"UCLA_BXDBXH_CARTILAGE_V2::ILM3140576",
"UCLA_BXDBXH_CARTILAGE_V2::ILM5670577",
"UCLA_BXDBXH_CARTILAGE_V2::ILM2070121",
"UCLA_BXDBXH_CARTILAGE_V2::ILM103990541",
"UCLA_BXDBXH_CARTILAGE_V2::ILM1190722",
"UCLA_BXDBXH_CARTILAGE_V2::ILM6590722",
"UCLA_BXDBXH_CARTILAGE_V2::ILM4200064",
"UCLA_BXDBXH_CARTILAGE_V2::ILM3140463"
]
}'
which should respond with a json response containing the raw binary string
for the png format and possibly another for the svg format.
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* Fix a few minor bugs left over from the integration of code from the
proof-of-concept code.
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* Add missing imports that are needed in the code.
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* Copy over code from the proof-of-concept implementation and clean it up a
little for the entry-point function for heatmap generation via the API
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* Add a function to generate the heatmaps for each chromosome into a single
plot.
|
|
* gn3/heatmaps.py: implement `process_traits_data_for_heatmap` function, that
will process the data into a form usable by heatmaps.
* tests/unit/test_heatmaps.py: check that the function processes the data into
the correct form.
|
|
* gn3/heatmaps.py: copy over function
* tests/unit/test_heatmaps.py: add tests
Copy function over from proof of concept and add some tests to ensure it
works as expected.
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* The heatmap generation does not fall cleanly within the computations or db
modules. This commit moves it to the higher level gn3 module.
|
|
* tests/unit/computations/test_heatmap.py: ordering is not longer provided as
a list of tuples; the ordering values are just a list of numbers now. This
commit updates the test to take this into consideration.
* tests/unit/computations/test_qtlreaper.py: the 'Chr' value if numeric, is
represented by an actual number, not a string. This commit updates the code
to take this into consideration.
|
|
* tests/unit/db/data/genotypes/genotype_sample1.geno: new file
Add a missing sample data file needed for unit tests.
|
|
* Add individual heatmaps
* Add dendograms
* Merge multiple heatmaps to single plot
Updated the proof of concept code to provide a sample of what is needed to
generate the appropriate heatmaps.
To generate the sample heatmaps, one can run something like:
env SQL_URI="mysql://webqtlout:webqtlout@127.0.0.1:3306/db_webqtl" \
python3 qtlfilesexport.py
assuming that the database can be accessed at 127.0.0.1:3306
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* Return a dict of values rather than list for the traits and chromosomes to
ease searching through the data.
|
|
* To ease sorting of data by numerical order down the line, sort the "Chr"
values by numerical order.
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* Change the id from 'T<n>' to simply '<n>' to ease sorting of the trait
results by numerical order rather than string order.
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* gn3/computations/heatmap.py: Fix ordering function
* tests/unit/computations/test_heatmap.py: update test
The order of the traits is important for the clustering algorithm, since the
clustering seems to use the distance of one trait from another to determine
how to order them.
This commit also gets rid of the xoffset argument that is not important to
the ordering, and was used in the older GN1 to determine how to draw the
clustering lines.
|
|
* gn3/computations/qtlreaper.py: Provide a function to organise the results by
trait for easier use down the line.
* tests/unit/computations/test_qtlreaper.py: provide a test to ensure that the
organising function works as expected.
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* The "Chr" value seems to be mostly a name of some sort, despite it being,
seemingly an number. This commit parses the "Chr" value as a string.
It also updates the tests to expec a string, rather than a number for "Chr"
values.
|
|
* gn3/computations/qtlreaper.py: handle exceptions
Sometimes, the values being parsed are plain strings and cannot be cast to
the float types. This commit handles that by casting only those values that
can be cast to float, and returning the others as strings.
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* Migrate the `web.webqtl.heatmap.Heatmap.getNearestMarker` function in GN1 to
GN3.
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* gn3/db/genotypes.py: parse genotype files
* tests/unit/db/test_genotypes.py: test parsing is correct
Add the overall genotype files parsing function and tests to check that the
parsing works as expected.
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* gn3/db/genotypes.py: parse data lines in file to genetic markers.
* tests/unit/db/test_genotypes.py: test that parsing works.
Add some tests to check that the parsing of the markers works as expected,
and add the code to actually parse the markers.
|
|
* gn3/db/genotypes.py: parse data header
* tests/unit/db/test_genotypes.py: check that header's parse works correctly.
Add tests to check that the parser works as expected. Add code to implement
the parsing and pass the tests.
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* gn3/db/genotypes.py: parse genotype labels
* tests/unit/db/test_genotypes.py: test that genotype labels are parsed
correctly
As part of parsing the genotype files into usable python data structures,
this commit adds a function to parse the label lines (beginning with "@")
into the appropriate values.
|
|
* Fix some linting errors and some minor bugs caught by the linter.
Move the `random_string` function to separate module for use in multiple
places in the code.
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* gn3/computations/heatmap.py: update function
* gn3/db/traits.py: new function
Remove extraneous data and arguments from the function.
- Load the genotype file
- Generate traits file
- Provide both raw traits data, and exported traits data in return
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* The number of the arguments to the function changed, and so the tests for
the function needed to be updated.
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* gn3/computations/qtlreaper.py: pass output files
* tests/unit/computations/data/qtlreaper/main_output_sample.txt: sample test
data
* tests/unit/computations/data/qtlreaper/permu_output_sample.txt: sample test
data
* tests/unit/computations/test_qtlreaper.py: add tests
Add code to parse the QTLReaper output data files.
|
|
heatmap_generation
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* gn3/computations/qtlreaper.py: Fix some bugs
* qtlfilesexport.py: Test out running rust-qtlreaper
Test out the qtlreaper interface code and fix some bugs caught in the
process.
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* gn3/db/genotypes.py: New module
* gn3/settings.py: Add new configuration variable
* qtlfilesexport.py: Test out new code
Add a module containing functions dealing with the genotype files.
Add a configuration variable to point to the location of the genotype files.
|
|
* Provide documentation on downloading and using the genotype files.
|
|
* README.md: update header: Traits ==> Trait
* gn3/computations/qtlreaper.py: update header: Traits ==> Trait
* qtlfilesexport.py: Choose only BXD strains
Rename the first column header from "Traits" to "Trait" to correspond with
what `rust-qtlreaper` expects.
Choose only the BXD strains for the proof-of-concept example - this helped
bring out the fact that the traits file SHOULD NOT contain a strain column
for a strain that does not exist in the genotype file in consideration.
If the traits file has a strain column which does not exist in the genotype
file, then `rust-qtlreaper` fails with a panic, since, from what I can tell,
it tries to get a value from the genotype file for the non-existent strain,
which results to a `None` type. Subsequent attempts at running an operation
on the `None` type lead to the panic.
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* Remove empty line at the end of the traits file
|
|
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* gn3/computations/heatmap.py: move `generate_traits_file` function to new
module
* gn3/computations/qtlreaper.py: new module to interface with the
`rust-qtlreaper` utility.
* gn3/settings.py: Provide setting for the path to the `rust-qtlreaper`
utility
* qtlfilesexport.py: Move `random_string` function to new module. Update to
use functions in new module.
Provide a module with functions to be used to interface with
`rust-qtlreaper`. This module essentially contains all the functions that
are needed to build the files needed for, and to run the qtlreaper utility.
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* gn3/heatmaps/heatmaps.py: document format of the traits file
To assist future developers, and development of the system, this commit
documents some of the hard-won knowledge about the operation of the system
to ease future development of the system.
The documentation, if good, might also help with future onboarding of new
developers to the system.
|
|
|
|
|
|
stratified permutations + covariates
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* As part of the development effort, this commit provides a proof-of-concept
as a reference for generating the traits data file.
It might be useful for verifying that the functions that are called are
working as is expected.
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* Seeing as not every requirement/feature has been migrated over at this time,
this commit just provides all the intermediate data representations in the
final return of the function for later use down the line.
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* Provide a function to export the given strains and traits data into a traits
file for use with `rust-qtlreaper`.
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* Rework the strains and values retrieval function to more closely correspond
to the working of the original code in GN1
|
|
|
|
Issue:
https://github.com/genenetwork/gn-gemtext-threads/blob/main/topics/gn1-migration-to-gn2/clustering.gmi
* guix.scm: (gn packages golang) ==> (gnu packages golang)
csvdiff has moved to upstream guix and been removed from latest
guix-bioinformatics.
|