Merge branch 'dev'

Author: nrminor

README.md

@@ -1,7 +1,7 @@
 # OneRoof: A Pipeline Prototype for Base-, Variant-, and Consensus-calling under One Proverbial Roof
-
 [![Nextflow](https://img.shields.io/badge/nextflow%20DSL2-%E2%89%A522.10.1-23aa62.svg)](https://www.nextflow.io/) [![run with docker](https://img.shields.io/badge/run%20with-docker-0db7ed?labelColor=000000&logo=docker)](https://www.docker.com/) [![run with singularity](https://img.shields.io/badge/run%20with-singularity-1d355c.svg?labelColor=000000)](https://sylabs.io/docs/) [![Docker CI](https://github.com/nrminor/oneroof/actions/workflows/docker-image.yaml/badge.svg)](https://github.com/nrminor/oneroof/actions/workflows/docker-image.yaml)
 
+
 - [Overview](#overview)
 - [Quick Start](#quick-start)
 - [Detailed Setup Instructions](#detailed-setup-instructions)
@@ -87,6 +87,10 @@ Most users should configure `oneroof` through the command line via the following
 | `--secondary` | None | Whether to turn on secondary alignments for each amplicon. Secondary alignments can increase depth at the cost of more reads potentially mapping to the wrong locations. By default, secondary alignments are off. |
 | `--min_consensus_freq` | 0.5 | The minimum required frequency of a variant base to be included in a consensu sequence. |
 | `--min_depth_coverage` | 20 | Minimum required depth of coverage to call a variant. |
+| `--min_variant_frequency` | 0.05 (illumina) or 0.10 (nanopore) | Minimum variant frequency to call a variant. |
+| `--meta_ref` | None | Dataset, either a local FASTA file or a pre-built dataset built by Sylph, to use for metagenomic profiling. Can download prebuilt ones here: https://github.com/bluenote-1577/sylph/wiki/Pre%E2%80%90built-databases. |
+| `--sylph_tax_db` | None | The taxonomic annotation for the sylph database specified with `--meta_ref`. The pipeline automaticially downloads the databases so only the identifier is needed here. |
+| `--nextclade_dataset` | None | The name of the dataset to run nextclade with. To see all dataset options run `nextclade dataset list --only-names`. |
 | `--results` | results/ | Where to place results. |
 | `--cleanup` | false | Whether to cleanup work directory after a successful run. |
 
@@ -136,4 +140,39 @@ Contributions, feature requests, improvement suggestions, and bug reports via Gi
 
 ## Citation
 
-Coming soon!
+> Lail, Andrew J., William C. Vuyk, Heather Machkovech, Nicholas R. Minor, Nura R. Hassan, Rhea Dalvie, Isla E. Emmen et al. “Amplicon sequencing of pasteurized retail dairy enables genomic surveillance of H5N1 avian influenza virus in United States cattle.” PloS one 20, no. 6 (2025): <https://doi.org/10.1371/journal.pone.0325203>.
+
+``` bibtex
+@article{Lail2025-xf,
+  title = "Amplicon sequencing of pasteurized retail dairy enables genomic surveillance of {H5N1} avian influenza virus in United States cattle",
+  author = "Lail, Andrew J and Vuyk, William C and Machkovech, Heather and Minor, Nicholas R and Hassan, Nura R and Dalvie, Rhea and Emmen, Isla E and Wolf, Sydney and Kalweit, Annabelle and Wilson, Nancy and Newman, Christina M and Tiburcio, Patrick Barros and Weiler, Andrea and Friedrich, Thomas C and O'Connor, David H",
+  journal = "PLoS One",
+  publisher = "Public Library of Science (PLoS)",
+  volume =  20,
+  number =  6,
+  pages = "e0325203",
+  month =  jun,
+  year =  2025,
+  copyright = "http://creativecommons.org/licenses/by/4.0/",
+  language = "en"
+}
+```
+
+> Kwon, Taeyong, Jessie D. Trujillo, Mariano Carossino, Heather M. Machkovech, Konner Cool, Eu Lim Lyoo, Gagandeep Singh et al. “Pathogenicity and transmissibility of bovine-derived HPAI H5N1 B3. 13 virus in pigs.” Emerging Microbes & Infections just-accepted (2025): <https://doi.org/10.1080/22221751.2025.2509742>.
+
+``` bibtex
+@ARTICLE{Kwon2025-yq,
+  title = "Pathogenicity and transmissibility of bovine-derived {HPAI} {H5N1} B3.13 virus in pigs",
+  author = "Kwon, Taeyong and Trujillo, Jessie D and Carossino, Mariano and Machkovech, Heather M and Cool, Konner and Lyoo, Eu Lim and Singh, Gagandeep and Kafle, Sujan and Elango, Shanmugasundaram and Vediyappan, Govindsamy and Wei, Wanting and Minor, Nicholas and Matias-Ferreyra, Franco S and Morozov, Igor and Gaudreault, Natasha N and Balasuriya, Udeni B R and Hensley, Lisa E and Diel, Diego G and Ma, Wenjun and Friedrich, Thomas C and Richt, Juergen A",
+  journal = "Emerg. Microbes Infect.",
+  publisher = "Informa UK Limited",
+  volume =  14,
+  number =  1,
+  pages = "2509742",
+  month =  dec,
+  year =  2025,
+  keywords = "Cattle; H5N1 genotype B3.13; highly pathogenic avian influenza; mammalian-like mutation; pathogenicity; pig; transmissibility",
+  copyright = "http://creativecommons.org/licenses/by-nc/4.0/",
+  language = "en"
+}
+```

bin/slack_alerts.py

@@ -6,7 +6,8 @@
 from pathlib import Path
 
 import pandas as pd
-import requests
+from slack_sdk import WebClient
+from slack_sdk.errors import SlackApiError
 
 
 def parse_command_line_args() -> argparse.Namespace:
@@ -52,6 +53,8 @@ def passing_samples(df: pd.DataFrame, coverage_threshold) -> tuple[str, int]:
             count += 1
             passing_line = f"{sample_id}: {proportion}\n"
             passing_message += passing_line
+    if passing_message == "":
+        passing_message = "No passing samples."
     return (passing_message, count)
 
 
@@ -65,6 +68,8 @@ def failing_samples(df: pd.DataFrame, coverage_threshold) -> str:
         if proportion < threshold:
             failing_line = f"{sample_id}: {proportion}\n"
             failing_message += failing_line
+    if failing_message == "":
+        failing_message = "No failing samples."
     return failing_message
 
 
@@ -112,36 +117,44 @@ def send_slack_notification(
     passing, count_passing = passing_samples(stats_df, coverage_threshold)
     failing = failing_samples(stats_df, coverage_threshold)
 
-    # creating the return message
-    message = (
-        f"Oneroof has finished successfully for experiment {run_label}, "
-        f"with {count_passing} samples passing. Below is a breakdown of "
-        f"which samples had greater than or equal to {coverage_threshold}X coverage."
-    )
-
-    results = f"PASSING\n-------\n{passing}\n\nFAILING\n-------\n{failing}"
-
-    complete_message = f"{message}\n```{results}```"
-
     user_id_list = get_user_ids()
     slack_token = get_slack_token()
 
+    block = [
+        {
+            "type": "section",
+            "text": {
+                "type": "plain_text",
+                "text": f"Oneroof has finished successfully for experiment {run_label}, with {count_passing} samples passing. Below is a breakdown of which samples had greater than or equal to {coverage_threshold}X coverage.",
+                "emoji": True,
+            },
+        },
+        {
+            "type": "section",
+            "fields": [
+                {
+                    "type": "mrkdwn",
+                    "text": f"*PASSING* \n ----------- \n {passing}  ",
+                },
+                {
+                    "type": "mrkdwn",
+                    "text": f"*FAILING* \n ----------- \n {failing}",
+                },
+            ],
+        },
+    ]
+
     for user_id in user_id_list:
-        resp = requests.post(
-            "https://slack.com/api/conversations.open",
-            headers={"Authorization": f"Bearer {slack_token}"},
-            json={"users": user_id},
-        )
-        channel_id = resp.json().get("channel", {}).get("id")
-        if not channel_id:
-            raise RuntimeError("Failed to open conversation.")
-
-        # Send the message
-        msg_resp = requests.post(
-            "https://slack.com/api/chat.postMessage",
-            headers={"Authorization": f"Bearer {slack_token}"},
-            json={"channel": channel_id, "text": complete_message},
-        )
+        try:
+            client = WebClient(token=slack_token)
+            # Send the message to a channel
+            response = client.chat_postMessage(
+                channel=user_id,  # or a user ID like "U12345678"
+                blocks=block,
+            )
+            print("Message sent successfully!")
+        except SlackApiError as e:
+            print(f"Error sending message: {e.response['error']}")
 
 
 def main() -> None:
@@ -151,7 +164,11 @@ def main() -> None:
     coverage_tsv_dir_path = args.input_tsv_dir
     coverage_depth = args.depth
 
-    send_slack_notification(run_label, coverage_tsv_dir_path, coverage_depth)
+    send_slack_notification(
+        run_label,
+        coverage_tsv_dir_path,
+        coverage_depth,
+    )
 
 
 if __name__ == "__main__":

conf/illumina.config

@@ -1,48 +1,50 @@
 params {
 
     // set platform
-    platform = "illumina"
+    platform                = "illumina"
 
     // Minimum variant frequency to call a variant.
-    min_variant_frequency = 0.05
+    min_variant_frequency   = 0.05
 
     // the minimum acceptable average quality for a given read
-    min_qual = 20
+    min_qual                = 20
 
     // split off ont results
     params.illumina_results = params.results + "/illumina"
 
     // Where to place the Nanopore workflow run command
-    run_command = params.illumina_results
+    run_command             = params.illumina_results
 
     // paired read merging
-    merged = params.illumina_results + "/01_merged_reads"
+    merged                  = params.illumina_results + "/01_merged_reads"
 
     // primer handling results
-    primer_handling = params.illumina_results + "/02_primer_handling"
-    respliced = params.primer_handling + "/01_respliced_primers"
-    complete_amplicons = params.primer_handling + "/02_complete_amplicons"
-    merged_by_sample = params.primer_handling + "/03_merged_by_sampe"
+    primer_handling         = params.illumina_results + "/02_primer_handling"
+    respliced               = params.primer_handling + "/01_respliced_primers"
+    complete_amplicons      = params.primer_handling + "/02_complete_amplicons"
+    merged_by_sample        = params.primer_handling + "/03_merged_by_sampe"
 
     // alignment results
-    alignment = params.illumina_results + "/03_alignments"
-    mosdepth = params.alignment + "/01_alignments"
-    cov_plots = params.alignment + "/02_coverage_plots"
+    alignment               = params.illumina_results + "/03_alignments"
+    mosdepth                = params.alignment + "/01_alignments"
+    cov_plots               = params.alignment + "/02_coverage_plots"
 
     // consensus results
-    consensus = params.illumina_results + "/04_consensus_seqs"
+    consensus               = params.illumina_results + "/04_consensus_seqs"
+
+    // metagenomic results 
+    metagenomics            = params.illumina_results + "/metagenomics"
 
     // variant results
-    variants = params.illumina_results + "/05_variants"
-    ivar = params.variants + "/01_ivar_tables"
-    vcf = params.variants + "/02_annotated_vcfs"
-    variant_tsv = params.variants + "/03_variant_tsv"
+    variants                = params.illumina_results + "/05_variants"
+    ivar                    = params.variants + "/01_ivar_tables"
+    vcf                     = params.variants + "/02_annotated_vcfs"
+    variant_tsv             = params.variants + "/03_variant_tsv"
 
     // qc results
-    qc = params.illumina_results + "/06_QC"
+    qc                      = params.illumina_results + "/06_QC"
 
     // phylo results
-    phylo = params.illumina_results + "07_phylo"
-    nextclade = params.phylo + "01_nextclade"
-
+    phylo                   = params.illumina_results + "07_phylo"
+    nextclade               = params.phylo + "01_nextclade"
 }

conf/nanopore.config

@@ -1,51 +1,56 @@
 params {
 
     // set platform
-    platform = "ont"
+    platform              = "ont"
 
     // Minimum variant frequency to call a variant.
     min_variant_frequency = 0.2
 
     // the minimum acceptable average quality for a given read
-    min_qual = 10
+    min_qual              = 10
 
     // split off ont results
-    params.ont_results = params.results + "/nanopore"
+    params.ont_results    = params.results + "/nanopore"
 
     // Where to place the Nanopore workflow run command
-    run_command = params.ont_results
+    run_command           = params.ont_results
 
     // basecalling and demultiplexing results
-    basecall_results = params.ont_results + "/01_basecalled_demuxed"
-    basecall_bams = params.basecall_results + "/bams"
-    basecall_fastqs = params.basecall_results + "/fastqs"
+    basecall_results      = params.ont_results + "/01_basecalled_demuxed"
+    basecall_bams         = params.basecall_results + "/bams"
+    basecall_fastqs       = params.basecall_results + "/fastqs"
 
     // primer handling results
-    primer_handling = params.ont_results + "/02_primer_handling"
-    respliced = params.primer_handling + "/01_respliced_primers"
-    complete_amplicons = params.primer_handling + "/02_complete_amplicons"
-    merged_by_sample = params.primer_handling + "/03_merged_by_sampe"
+    primer_handling       = params.ont_results + "/02_primer_handling"
+    respliced             = params.primer_handling + "/01_respliced_primers"
+    complete_amplicons    = params.primer_handling + "/02_complete_amplicons"
+    merged_by_sample      = params.primer_handling + "/03_merged_by_sampe"
 
     // alignment results
-    alignment = params.ont_results + "/03_alignments"
-    mosdepth = params.alignment + "/01_mosdepth"
-    cov_plots = params.alignment + "/02_coverage_plots"
+    alignment             = params.ont_results + "/03_alignments"
+    mosdepth              = params.alignment + "/01_mosdepth"
+    cov_plots             = params.alignment + "/02_coverage_plots"
 
     // consensus results
-    consensus = params.ont_results + "/04_consensus_seqs"
+    consensus             = params.ont_results + "/04_consensus_seqs"
+
+    // metagenomic results 
+    metagenomics          = params.ont_results + "/metagenomics"
+
+    //haplyotyping results 
+    haplotyping           = params.ont_results + "/haplotyping"
 
     // variant results
-    variants = params.ont_results + "/05_variants"
-    ivar = params.variants + "/01_ivar_tables"
-    vcf = params.variants + "/02_annotated_vcfs"
-    variant_tsv = params.variants + "/03_variant_tsv"
-    haplo = params.variants + "/04_haplotypes"
+    variants              = params.ont_results + "/05_variants"
+    ivar                  = params.variants + "/01_ivar_tables"
+    vcf                   = params.variants + "/02_annotated_vcfs"
+    variant_tsv           = params.variants + "/03_variant_tsv"
+    haplo                 = params.variants + "/04_haplotypes"
 
     // qc results
-    qc = params.ont_results + "/06_QC"
+    qc                    = params.ont_results + "/06_QC"
 
     // phylo results
-    phylo = params.ont_results + "/07_phylo"
-    nextclade = params.phylo + "/01_nextclade"
-
+    phylo                 = params.ont_results + "/07_phylo"
+    nextclade             = params.phylo + "/01_nextclade"
 }