From ed23b6b0c56f4fb2184fff1fd528f86a45f4962d Mon Sep 17 00:00:00 2001
From: Eric Nawrocki <nawrocke@ncbi.nlm.nih.gov>
Date: Thu, 15 Aug 2024 17:56:52 +0100
Subject: [PATCH] Removes SS_cons gaps from cmemit -c | cmalign alignment This
 is rare but occurs for some families if cmalign inserts a consensus nt
 followed by a nearby deletion. Examples for Rfam 15.0 are RF00023 and
 RF03057.

---
 Rfam/Lib/Bio/Rfam/QC.pm | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/Rfam/Lib/Bio/Rfam/QC.pm b/Rfam/Lib/Bio/Rfam/QC.pm
index 63c3ef0c..b6324c2a 100644
--- a/Rfam/Lib/Bio/Rfam/QC.pm
+++ b/Rfam/Lib/Bio/Rfam/QC.pm
@@ -261,6 +261,10 @@ sub checkCMFormat {
     chomp $line;
     if($line =~ /^#=GC\s+SS_cons\s+(\S+)/) {
       $cm_ss_cons = $1;
+      # remove any gaps, this is rare but gaps can occur if cmalign thinks a consensus position
+      # should be inserted followed by a nearby deletion (e.g. RF00023 and RF03057 for Rfam
+      # 15.0 with infernal 1.1.5)
+      $cm_ss_cons =~ s/\.//g;
     }
     else {
       warn "Failed to read SS_cons from alignment created with command $cmd\n";