From c6eeafe81303661b2723e6256e3ee521a4fee7a2 Mon Sep 17 00:00:00 2001
From: LiaOb21 <liaobinu@gmail.com>
Date: Thu, 22 Feb 2024 18:00:21 +0000
Subject: [PATCH] adding config used for testing the test dataset

---
 config/config_test.yaml | 67 +++++++++++++++++++++++++++++++++++++++++
 1 file changed, 67 insertions(+)
 create mode 100644 config/config_test.yaml

diff --git a/config/config_test.yaml b/config/config_test.yaml
new file mode 100644
index 0000000..3a4f78d
--- /dev/null
+++ b/config/config_test.yaml
@@ -0,0 +1,67 @@
+# config.yaml
+
+# Path to hifi reads
+hifi_path: "resources/raw_hifi/"
+
+# Path to hic reads
+hic_path: "resources/raw_hic/"
+
+# Customisable parameters for nanoplot
+nanoplot:
+  t: 4 # number of threads
+
+# Customisable parameters for fastp 
+fastp:
+  t: 4 # number of threads
+  cut_window_size: 4
+  cut_mean_quality: 20
+  optional_params: 
+    "--detect_adapter_for_pe": "true"
+    "--cut_front": "true"
+    "--cut_tail": "true"
+    "-p": "false"
+
+# Customisable parameters for kmc
+kmc:
+  k: 21 # kmer size, it will be the same used for genomescope2
+  t: 4 # number of threads
+  ci: 1 # exclude k-mers occurring less than <value> times (default: 2)
+  cs: 100000 #maximal value of a counter (default: 255)
+
+# Customisable parameters for kmc_tools transform
+kmc_tools:
+  cx: 100000 # exclude k-mers occurring more of than <value> times
+
+# Customisable parameters for oatk
+oatk:
+  k: 1001 # kmer size [1001]
+  c: 150 #  minimum kmer coverage [3]
+  t: 4  # number of threads [1]
+  m: "resources/oatkDB/dikarya_mito.fam" # mitochondria gene annotation HMM profile database [NULL]
+  optional_params: 
+    "-p": "" # to use for species that have a plastid db
+
+# Customisable parameters for minimap2
+minimap2:
+  t: 4  # number of threads
+
+# Customisable parameters for hifiasm
+hifiasm:
+  t: 4  # number of threads
+  optional_params: 
+    "-f": "0" # used for small datasets
+    "-l": "" # purge level. 0: no purging; 1: light; 2/3: aggressive [0 for trio; 3 for unzip]
+    "--ul": "resources/raw_ont/ont_test_SPSC01_SRR27947616_PRJNA1075684.fastq.gz"
+    "--h1": ""
+    "--h2": ""
+
+# Customisable parameters for arima mapping pipeline:
+arima:
+  MAPQ_FILTER: 10
+  CPU: 4
+
+# Customisable parameters for yahs
+yahs:
+  o: "hifiasm_p_purged_yahs" # output prefix
+  optional_params: 
+    "-e": "GATC" # you can specify the restriction enzyme(s) used by the Hi-C experiment