Merge pull request #44 from sanjaynagi/move-ag3-af1-17-05-24

move ag3/af1 to after species selection

AnoPrimer/AnoPrimer.py

@@ -10,19 +10,16 @@
 from matplotlib import patches
 from plotly.subplots import make_subplots
 
-ag3 = malariagen_data.Ag3(url="gs://vo_agam_release/", pre=True)
-af1 = malariagen_data.Af1(url="gs://vo_afun_release/", pre=True)
-
 
 def retrieve_data_resource(species):
     assert species in [
         "gambiae_sl",
         "funestus",
     ], f"species {species} not recognised, please use 'gambiae_sl' or 'funestus'"
     if species == "gambiae_sl":
-        data_resource = ag3
+        data_resource = malariagen_data.Ag3(url="gs://vo_agam_release/", pre=True)
     elif species == "funestus":
-        data_resource = af1
+        data_resource = malariagen_data.Af1(url="gs://vo_afun_release/", pre=True)
     return data_resource
 
 
@@ -605,7 +602,9 @@ def plot_primer_locs(
     # plot the legend
     plt.legend(handles=handles, loc=legend_loc)
     if out_dir:
-        fig.savefig(f"{out_dir}/{assay_name}_primer_locs.png", dpi=300)
+        fig.savefig(
+            f"{out_dir}/{assay_name}_primer_locs.png", dpi=300, bbox_inches="tight"
+        )
 
 
 def gget_blat_genome(primer_df, assay_type, assembly="anoGam3"):
@@ -1256,21 +1255,24 @@ def check_my_oligo(
     contig = contig.replace("chr", "")
     region_span = f"{contig}:{start}-{end}"
     print("plotting frequencies in ag3 data")
+
     fig = plot_sequence_frequencies(
+        data_resource=malariagen_data.Ag3(url="gs://vo_agam_release/", pre=True),
         region=region_span,
         sample_sets=sample_sets,
         sample_query=sample_query,
         width=width,
         height=height,
     )
+    return fig
 
 
 def plot_sequence_frequencies(
-    region, sample_sets=None, sample_query=None, width=700, height=400
+    data_resource, region, sample_sets=None, sample_query=None, width=700, height=400
 ):
     """Retrieve frequencies"""
 
-    snps = ag3.snp_calls(
+    snps = data_resource.snp_calls(
         region=region, sample_sets=sample_sets, sample_query=sample_query
     )
     ref_alt_arr = snps["variant_allele"].compute().values.astype(str)

docs/AnoPrimer-docs/_toc.yml

@@ -5,4 +5,5 @@ parts:
   chapters:
     - file: notebooks/api
     - file: notebooks/notebooks
+    - file: notebooks/data_access
     - file: notebooks/troubleshooting_contributing

notebooks/AnoPrimer-long.ipynb

@@ -8,7 +8,7 @@
         "id": "view-in-github"
       },
       "source": [
-        "<a href=\"https://colab.research.google.com/github/sanjaynagi/AgamPrimer/blob/main/notebooks/AgamPrimer-long.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
+        "<a href=\"https://colab.research.google.com/github/sanjaynagi/AnoPrimer/blob/main/notebooks/AnoPrimer-long.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
       ]
     },
     {
@@ -25,24 +25,40 @@
       "outputs": [],
       "source": [
         "# First, install some packages we require\n",
-        "! pip install AgamPrimer -q "
+        "! pip install AnoPrimer -q "
       ]
     },
     {
       "cell_type": "code",
-      "execution_count": null,
+      "execution_count": 2,
       "id": "528ae85d",
       "metadata": {
         "id": "528ae85d"
       },
-      "outputs": [],
+      "outputs": [
+        {
+          "ename": "ModuleNotFoundError",
+          "evalue": "No module named 'AnoPrimer'",
+          "output_type": "error",
+          "traceback": [
+            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
+            "\u001b[0;31mModuleNotFoundError\u001b[0m                       Traceback (most recent call last)",
+            "Cell \u001b[0;32mIn[2], line 2\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[38;5;66;03m# Import libraries \u001b[39;00m\n\u001b[0;32m----> 2\u001b[0m \u001b[38;5;28;01mimport\u001b[39;00m \u001b[38;5;21;01mAnoPrimer\u001b[39;00m\n\u001b[1;32m      3\u001b[0m \u001b[38;5;28;01mimport\u001b[39;00m \u001b[38;5;21;01mmalariagen_data\u001b[39;00m\n\u001b[1;32m      4\u001b[0m \u001b[38;5;28;01mimport\u001b[39;00m \u001b[38;5;21;01mprimer3\u001b[39;00m\n",
+            "\u001b[0;31mModuleNotFoundError\u001b[0m: No module named 'AnoPrimer'"
+          ]
+        }
+      ],
       "source": [
         "# Import libraries \n",
-        "import AgamPrimer\n",
+        "import AnoPrimer\n",
         "import malariagen_data\n",
         "import primer3\n",
         "import pandas as pd\n",
-        "import numpy as np"
+        "import numpy as np\n",
+        "\n",
+        "import google.auth\n",
+        "\n",
+        "credentials, project = google.auth.default()"
       ]
     },
     {
@@ -53,11 +69,11 @@
         "id": "47bde61a"
       },
       "source": [
-        "##**[AgamPrimer](https://github.com/sanjaynagi/AgamPrimer): Primer design considering genomic variation in *Anopheles gambiae* s.l and *Anopheles funestus***   \n",
+        "##**[AnoPrimer](https://github.com/sanjaynagi/AnoPrimer): Primer design considering genomic variation in *Anopheles gambiae* s.l and *Anopheles funestus***   \n",
         "**Author**: [Sanjay Curtis Nagi](https://sanjaynagi.github.io/)    \n",
         "**Email**: sanjay.nagi@lstmed.ac.uk  \n",
         "\n",
-        "Welcome to the AgamPrimer notebook. This is the long version of the notebook which designs primers in steps. Alternatively, there is a all-in-one function in a [short version of the notebook](https://colab.research.google.com/github/sanjaynagi/AgamPrimer/blob/main/notebooks/AgamPrimer-short.ipynb).\n",
+        "Welcome to the AnoPrimer notebook. This is the long version of the notebook which designs primers in steps. Alternatively, there is a all-in-one function in a [short version of the notebook](https://colab.research.google.com/github/sanjaynagi/AnoPrimer/blob/main/notebooks/AnoPrimer-short.ipynb).\n",
         "\n",
         "We would like to design primers for PCR applications, such as genotyping or gene expression (qPCR). However, single nucleotide polymorphisms (SNPs) in primer binding sites can result in differences or failures in PCR amplification, referred to as null alleles. \n",
         "\n",
@@ -128,9 +144,9 @@
       },
       "outputs": [],
       "source": [
-        "data_resource = AgamPrimer.retrieve_data_resource(species)\n",
+        "data_resource = AnoPrimer.retrieve_data_resource(species)\n",
         "# Connect to the malariagen_data ag3 API\n",
-        "contig, target = AgamPrimer.check_and_split_target(species=species, target=primer_target, assay_type=assay_type)\n",
+        "contig, target = AnoPrimer.check_and_split_target(species=species, target=primer_target, assay_type=assay_type)\n",
         "genome_seq = data_resource.genome_sequence(region=contig)\n",
         "print(f\"Our genome sequence for {contig} is {genome_seq.shape[0]} bp long\")"
       ]
@@ -143,7 +159,7 @@
         "id": "144a37ce"
       },
       "source": [
-        "Now we need to extract the bit of sequence we need. We will use functions in the [AgamPrimer](https://pypi.org/project/AgamPrimer/) package."
+        "Now we need to extract the bit of sequence we need. We will use functions in the [AnoPrimer](https://pypi.org/project/AnoPrimer/) package."
       ]
     },
     {
@@ -159,7 +175,7 @@
       },
       "outputs": [],
       "source": [
-        "target_sequence, gdna_pos, seq_parameters = AgamPrimer.prepare_sequence(\n",
+        "target_sequence, gdna_pos, seq_parameters = AnoPrimer.prepare_sequence(\n",
         "                                                                species=species,\n",
         "                                                                target=target,\n",
         "                                                                assay_type=assay_type,\n",
@@ -236,7 +252,7 @@
         "        # In the same format as above                       \n",
         "    }\n",
         "\n",
-        "primer_parameters = AgamPrimer.primer_params(primer_parameters=primer_parameters, assay_type=assay_type, n_primer_pairs=n_primer_pairs, amplicon_size_range=amplicon_size_range) ## adds some parameters depending on assay type"
+        "primer_parameters = AnoPrimer.primer_params(primer_parameters=primer_parameters, assay_type=assay_type, n_primer_pairs=n_primer_pairs, amplicon_size_range=amplicon_size_range) ## adds some parameters depending on assay type"
       ]
     },
     {
@@ -285,7 +301,7 @@
       },
       "outputs": [],
       "source": [
-        "AgamPrimer.primer3_run_statistics(primer_dict, assay_type)"
+        "AnoPrimer.primer3_run_statistics(primer_dict, assay_type)"
       ]
     },
     {
@@ -312,7 +328,7 @@
       },
       "outputs": [],
       "source": [
-        "primer_df = AgamPrimer.primer3_to_pandas(primer_dict=primer_dict, assay_type=assay_type)\n",
+        "primer_df = AnoPrimer.primer3_to_pandas(primer_dict=primer_dict, assay_type=assay_type)\n",
         "primer_df"
       ]
     },
@@ -474,7 +490,7 @@
       },
       "outputs": [],
       "source": [
-        "results_dict = AgamPrimer.plot_primer_snp_frequencies(\n",
+        "results_dict = AnoPrimer.plot_primer_snp_frequencies(\n",
         "    species=species,\n",
         "    primer_df=primer_df,\n",
         "    gdna_pos=gdna_pos,\n",
@@ -511,7 +527,7 @@
       },
       "outputs": [],
       "source": [
-        "AgamPrimer.plot_primer_locs(\n",
+        "AnoPrimer.plot_primer_locs(\n",
         "    species=species,\n",
         "    primer_df=primer_df, \n",
         "    primer_res_dict=results_dict,\n",
@@ -567,7 +583,7 @@
       "outputs": [],
       "source": [
         "if species == 'gambiae_sl':\n",
-        "    blat_result_df = AgamPrimer.gget_blat_genome(\n",
+        "    blat_result_df = AnoPrimer.gget_blat_genome(\n",
         "        primer_df, \n",
         "        assay_type, \n",
         "        assembly='anoGam3'\n",
@@ -596,7 +612,7 @@
       "outputs": [],
       "source": [
         "#view help for function\n",
-        "AgamPrimer.designPrimers?"
+        "AnoPrimer.designPrimers?"
       ]
     },
     {
@@ -613,7 +629,7 @@
       },
       "outputs": [],
       "source": [
-        "# primer_df, blat_df = AgamPrimer.designPrimers(\n",
+        "# primer_df, blat_df = AnoPrimer.designPrimers(\n",
         "#     species=species,\n",
         "#     assay_type='gDNA primers + probe',\n",
         "#     target='X:2422652' ,\n",
@@ -651,7 +667,7 @@
         "\n",
         "####**Future development**\n",
         "\n",
-        "Any contributions or suggestions on how we can improve this notebook are more than welcome. Please [email](mailto:sanjay.nagi@lstmed.ac.uk) or log an [issue on github](https://github.com/sanjaynagi/primerDesignAg/issues). This notebook and source code for AgamPrimer are located here - https://github.com/sanjaynagi/AgamPrimer/    \n",
+        "Any contributions or suggestions on how we can improve this notebook are more than welcome. Please [email](mailto:sanjay.nagi@lstmed.ac.uk) or log an [issue on github](https://github.com/sanjaynagi/primerDesignAg/issues). This notebook and source code for AnoPrimer are located here - https://github.com/sanjaynagi/AnoPrimer/    \n",
         "\n",
         "\\\n",
         "####**References**\n",
@@ -684,7 +700,7 @@
       "name": "python",
       "nbconvert_exporter": "python",
       "pygments_lexer": "ipython3",
-      "version": "3.7.13 (default, Mar 29 2022, 02:18:16) \n[GCC 7.5.0]"
+      "version": "3.10.12"
     },
     "vscode": {
       "interpreter": {