diff --git a/minos/scripts/metric_oddities.py b/minos/scripts/metric_oddities.py
index 7e4029c..47ec421 100644
--- a/minos/scripts/metric_oddities.py
+++ b/minos/scripts/metric_oddities.py
@@ -1,3 +1,4 @@
+import sys
 import csv
 from collections import Counter
 
@@ -6,18 +7,33 @@
 
 
 class MetricOddityParser:
-	def __init__(self, metric_file, oddities, gene_filter=None):
-		self.table = Counter({oddity: 0 for oddity in oddities})
-		self.metric_file = metric_file
-		self.gene_filter = gene_filter
-	def run(self):
-		for row in csv.DictReader(open(self.metric_file), delimiter="\t"):
-			if gene_filter is None or row["tid"] in gene_filter:
-				counts = [oddity for oddity in self.table if eval(oddity.format(**row))]
-				self.table.update(counts)
-
-		for oddity, count in self.table.items():
-			print(oddity.replace("{", "").replace("}", ""), count, sep="\t")
-
 
+	def parse_metrics(self, metric_file, transcript_filter=None):
+		for row in csv.DictReader(open(metric_file), delimiter="\t"):
+			if transcript_filter is None or row["tid"] in transcript_filter:
+				counts = [oddity for oddity in self.oddities if eval(oddity.format(**row))]
+				source = row["original_source"]
+				self.data.setdefault(
+					source,
+					Counter({oddity: 0 for oddity in self.oddities})
+				).update(counts)
 				
+	def __init__(self, metric_file, oddities, transcript_filter=None):
+		self.data = dict()
+		self.oddities = oddities
+		self.parse_metrics(metric_file, transcript_filter=transcript_filter)
+
+	def write_table(self, collapse=True, stream=sys.stdout):
+		# collapse: True for loci, False for subloci and monoloci
+		header = list(self.data.keys()) if not collapse else ["counts"]
+		print("metric", *header, sep="\t", file=stream, flush=True)
+		for oddity in self.oddities:
+			total = 0
+			row = list()
+			for source, counts in self.data.items():
+				total += counts[oddity]
+				if not collapse:
+					row.append(counts[oddity])
+			if collapse:
+				row.append(total)
+			print(oddity.replace("{", "").replace("}", ""), *row, sep="\t", file=stream, flush=True)
diff --git a/minos/zzz/minos_run.smk b/minos/zzz/minos_run.smk
index 653d733..816ec6b 100644
--- a/minos/zzz/minos_run.smk
+++ b/minos/zzz/minos_run.smk
@@ -167,6 +167,9 @@ rule all:
 		os.path.join(config["outdir"], "mikado.subloci.gff3"),
 		os.path.join(config["outdir"], "mikado.monoloci.gff3"),
 		os.path.join(config["outdir"], "mikado.loci.gff3"),
+		os.path.join(RESULTS_DIR, "mikado.subloci.metrics_oddities.tsv"),
+		os.path.join(RESULTS_DIR, "mikado.monoloci.metrics_oddities.tsv"),
+		os.path.join(RESULTS_DIR, "mikado.loci.metrics_oddities.tsv"),
 		[
 			os.path.join(config["outdir"], POST_PICK_PREFIX + suffix)
 			for suffix in {".gff", ".proteins.fasta", ".table.txt", ".cds.fasta", ".cds.fasta.lengths", ".cdna.fasta"}
@@ -876,21 +879,28 @@ rule minos_generate_final_table:
 
 rule minos_collate_metric_oddities:
 	input:
-		loci = os.path.join(config["outdir"], "mikado.loci.gff3"),
-		subloci = os.path.join(config["outdir"], "mikado.subloci.gff3"),
-		monoloci = os.path.join(config["outdir"], "mikado.monoloci.gff3"),	
+		loci = os.path.join(config["outdir"], "mikado.loci.metrics.tsv"),
+		subloci = os.path.join(config["outdir"], "mikado.subloci.metrics.tsv"),
+		monoloci = os.path.join(config["outdir"], "mikado.monoloci.metrics.tsv"),
+		old_new_rel = rules.minos_create_release_gffs.output[2],
 		final_table = rules.minos_generate_final_table.output.final_table
 	output:
-		rules.minos_final_sanity_check.output[0] + ".metric_oddities.tsv"
+		os.path.join(config["outdir"], "results", "mikado.loci.metrics_oddities.tsv"),
+		os.path.join(config["outdir"], "results", "mikado.subloci.metrics_oddities.tsv"),
+		os.path.join(config["outdir"], "results", "mikado.monoloci.metrics_oddities.tsv")
 	run:
+		import csv
 		from minos.scripts.metric_oddities import MetricOddityParser
-		tx2gene = {row[1]: row[0] for row in csv.reader(open(input.final_table), delimiter="\t") if not row[0].startswith("#")}
-		release_genes = set(tx2gene.values())
+		#tx2gene = {row[1]: row[0] for row in csv.reader(open(input.final_table), delimiter="\t") if not row[0].startswith("#")}
+		#release_genes = set(tx2gene.values())
+		release_transcripts = {row[0] for row in csv.reader(open(input.final_table), delimiter="\t") if not row[0].startswith("#")}
+		transcript_filter = {row[3] for row in csv.reader(open(input.old_new_rel), delimiter="\t") if not row[0].startswith("#") and row[1] in release_transcripts}
 		with open(output[0], "w") as loci_oddities_out:
-			MetricOddityParser(input.loci, config["report_metric_oddities"], release_genes).run(stream=loci_oddities_out)
-		
-	
-
+			MetricOddityParser(input[0], config["report_metric_oddities"], transcript_filter=transcript_filter).write_table(collapse=True, stream=loci_oddities_out)
+		with open(output[1], "w") as subloci_oddities_out:
+			MetricOddityParser(input[1], config["report_metric_oddities"]).write_table(collapse=False, stream=subloci_oddities_out)
+		with open(output[2], "w") as monoloci_oddities_out:
+			MetricOddityParser(input[2], config["report_metric_oddities"]).write_table(collapse=False, stream=monoloci_oddities_out)
 
 rule split_proteins_prepare:
 	input: