From 1c43e124c6c5362b14d61dc2914eb78d2b26906a Mon Sep 17 00:00:00 2001
From: Fredrik Boulund <fredrik.boulund@ki.se>
Date: Tue, 16 Nov 2021 11:04:21 +0100
Subject: [PATCH] Update Kraken2, fastp, kajiu, bbmap

---
 CHANGELOG.md                          | 10 ++++++++--
 config.yaml                           |  1 +
 envs/stag-mwc.yaml                    |  6 +++---
 rules/taxonomic_profiling/kraken2.smk |  2 ++
 4 files changed, 14 insertions(+), 5 deletions(-)

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 7c0b6ca..5698243 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -18,8 +18,8 @@ situations.
 ### Added
 - Biobakery update: updated MetaPhlAn and HUMAnN to version 3 as well as introducing 
   StrainPhlAn3 for strain-level genomics.
-- Added $TMPDIR variable which can be specified in config.yaml. It is normally not required
-  to specify $TMPDIR but might be necessary to run HUMAnN due to large intermediary files.
+- Added `$TMPDIR` variable which can be specified in config.yaml. It is normally not required
+  to specify `$TMPDIR` but might be necessary to run HUMAnN due to large intermediary files.
 - New internal StaG feature to better handle user messages and defer them for printing after
   the workflow finished execution so they don't get lost in the verbose log printout from
   Snakemake.
@@ -27,6 +27,8 @@ situations.
   `keep_` flags in the config file. Currently available for Quality control
   (fastp), host removal (kraken2), taxonomic profiling with Kraken2 and
   MetaPhlAn3.
+- Automatic builds of Singularity images using Github actions. Images get uploaded to
+  the Github Package repository at `ghcr.io/ctmrbio/stag-mwc:<tag>-<branch>`.
 
 ### Fixed
 - Updated pandas to 1.2.1 to fix issue with `preprocessing_summary.py` failing.
@@ -38,6 +40,10 @@ situations.
 - Limited job allocations to one node in `ctmr_gandalf` cluster config.
 
 ### Changed
+- Updated Kraken2 to 2.1.2 and added `--minimum-hit-groups` argument in config file.
+- Updated fastp to 0.23.0.
+- Updated Kaiju to 1.8.2.
+- Updated BBMap to 38.93.
 
 ### Removed
 
diff --git a/config.yaml b/config.yaml
index dc16269..3f9c206 100644
--- a/config.yaml
+++ b/config.yaml
@@ -88,6 +88,7 @@ kaiju:
 kraken2:
     db: ""                   # [Required] Path to Kraken2 DB folder
     confidence: 0.1          # Kraken2 confidence score, float in [0,1]
+    minimum_hit_groups: 2    # Integer, default 2
     extra: ""                # Extra command line arguments for kraken2 (do not add/change output files)
     keep_kraken: False       # Keep the kraken output files
     keep_kreport: True       # Keep the kreport output files
diff --git a/envs/stag-mwc.yaml b/envs/stag-mwc.yaml
index b32132a..e197e2e 100644
--- a/envs/stag-mwc.yaml
+++ b/envs/stag-mwc.yaml
@@ -5,9 +5,9 @@ channels:
     - defaults
 dependencies:
     - python =3.7
-    - fastp =0.20.0
-    - bbmap =38.68
-    - kaiju =1.7.2
+    - fastp =0.23.0
+    - bbmap =38.93
+    - kaiju =1.8.2
     - kraken2 =2.1.2
     - bracken =2.5
     - krona =2.7
diff --git a/rules/taxonomic_profiling/kraken2.smk b/rules/taxonomic_profiling/kraken2.smk
index f956b37..9b620d1 100644
--- a/rules/taxonomic_profiling/kraken2.smk
+++ b/rules/taxonomic_profiling/kraken2.smk
@@ -93,12 +93,14 @@ rule kraken2:
     params:
         db=kraken2_config["db"],
         confidence=kraken2_config["confidence"],
+        minimum_hit_groups=kraken2_config["minimum_hit_groups"],
         extra=kraken2_config["extra"],
     shell:
         """
         kraken2 \
             --db {params.db} \
             --confidence {params.confidence} \
+            --minimum-hit-groups {params.minimum_hit_groups} \
             --threads {threads} \
             --output {output.kraken} \
             --report {output.kreport} \