1. ncRNA

非编码RNA(Non-coding RNA, ncRNA) 包括rRNA，tRNA，snRNA，snoRNA 和microRNA 等不编码蛋白质的RNA，它们转录后直接在RNA 水平上就能行使各自的生物学功能，并不需要翻译成蛋白质。

2. 软件

tRNA注释

一般用tRNAscan-SE，老牌软件。

rRNA注释

可用RNAMMER注释。但需要edu邮箱下载。

或者用barrnap，开源软件，也非常好用。

其他ncRNA注释

Infernal软件+Rfam数据库是注释其他复杂ncRNA的常用组合，它的结果同时包含了tRNA，rRNA，snRNA，snoRNA和miRNA等。

对比以上两款软件的tRNA和rRNA结果，Infernal得到的结果差不太多。所以如果不重点研究ncRNA，用这个组合也可。

3. 注释

tRNA

以tRNAscan为例：

tRNAscan-SE genome.fa -o tRNA.out -f tRNA.ss -m tRNA.stats

tRNA.stats文件中包含了tRNA预测的结果统计。

image.png

可通过脚本将tRNA.out整理为gff3。

perl convert_tRNAScanSE_to_gff3.pl -i tRNA.out >tRNAscan.gff3

PS：convert_tRNAScanSE_to_gff3.pl 可参考：github: https://github.com/jorvis/biocode/blob/master/gff/convert_tRNAScanSE_to_gff3.pl

rRNA

以barrnap为例：

barrnap-master/bin/barrnap \
    --kingdom euk \   #euk bac arc mito
    --threads 10 \
    --outseq rRNA.fa \ #rRNA结果序列
    genome.fa >rRNA.gff3  #rRNA gff结果

PS：不建议用conda安装使用，会报错。如[barrnap] ERROR: nhmmer failed to run - Error: Invalid alphabet type in target for nhmmer. Expect DNA or RNA.

得到的结果可进一步统计各核糖体亚基数目。

• 细菌bacteria (5S,23S,16S)
• 古菌archaea (5S,5.8S,23S,16S)
• 多细胞生物线粒体metazoan mitochondria (12S,16S)
• 真核生物eukaryotes (5S,5.8S,28S,18S)

snRNA、miRNA等

首次下载安装infernal和Rfam数据库。infernal使用conda方便，Rfam数据库重点是Rfam.cm和Rfam.clanin的下载：

# install the package with conda 
conda install -c bioconda infernal

#  download Rfam data 
wget ftp://ftp.ebi.ac.uk/pub/databases/Rfam/CURRENT/Rfam.cm.gz
gunzip Rfam.cm.gz
wget ftp://ftp.ebi.ac.uk/pub/databases/Rfam/CURRENT/Rfam.clanin

# cmpress 索引
cmpress Rfam.cm

然后使用cmscan注释ncRNA：

genome_total=400000000  #基因组大小
CMnumber=`grep "ACC" /database/Rfam/Rfam.cm|wc -l`
Z=`echo $genome_total*2*$CMnumber/1000000 |bc`
#echo $Z

cmscan -Z $Z --cut_ga --rfam --nohmmonly --tblout genome.tblout  \
  --fmt 2 --cpu 30 --clanin /database/Rfam/Rfam.clanin \
  /database/Rfam/Rfam.cm genome.fa > genome.cmscan

主要有两个结果文件genome.cmscan（比对结果）和genome.tblout（table格式）。要想得到进一步分类统计结果，需要做一些处理。

4. snRNA、miRNA等结果的统计

首先，提取必须的列和非重叠区域或重叠区域中得分高的区域。

awk 'BEGIN{OFS="\t";}{if(FNR==1) print "target_name\taccession\tquery_name\tquery_start\tquery_end\tstrand\tscore\tEvalue"; if(FNR>2 && $20!="=" && $0!~/^#/) print $2,$3,$4,$10,$11,$12,$17,$18; }' my-genome.tblout >genome.tblout.final.xls

image.png

然后，需要根据Rfam数据库的注释来对结果进行分类。Rfam的注释可从官网拷贝（命名tmp)：https://rfam.xfam.org/search/type

image.png

image.png

image.png

# 拆分第三列
less tmp | awk 'BEGIN {FS=OFS="\t"}{split($3,x,";");class=x[2];print $1,$2,$3,$4,class}' > rfam_anno.txt

最后统计各类ncRNA注释结果。

 awk 'BEGIN{OFS=FS="\t"}ARGIND==1{a[$2]=$5;}ARGIND==2{type=a[$1]; if(type=="") type="Others"; count[type]+=1;}END{for(type in count) print type, count[type];}' rfam_anno.txt genome.tblout.final.xls

统计结果如下：

 riboswitch 1
 tRNA   813
 ribozyme   1
Others  217
 miRNA  1948
 antisense  7
 rRNA   1451
 sRNA   1
 snRNA  667

也可将注释结果整理为gff3文件：

perl infernal-tblout2gff.pl --cmscan --fmt2 genome.tblout >genome.infernal.ncRNA.gff3

附infernal-tblout2gff.pl

#!/usr/bin/env perl
# infernal-tblout2gff.pl: convert cmsearch or cmscan tblout files to GFF format

# an important point of above ref:
# "Start is always less than or equal to end"
# EPN, Fri Jun  7 11:07:38 2019
# 
#
use strict;
use warnings;
use Getopt::Long;

my $in_tblout  = "";   # name of input tblout file

my $usage;
$usage  = "infernal-tblout2gff.pl\n\n";
$usage .= "Usage:\n\n";
$usage .= "infernal-tblout2gff.pl [OPTIONS] <cmsearch tblout file>\n\tOR\n";
$usage .= "infernal-tblout2gff.pl --cmscan [OPTIONS] <cmscan tblout file>\n\tOR\n";
$usage .= "infernal-tblout2gff.pl --cmscan --fmt2 [OPTIONS] <cmscan --fmt 2 tblout file>\n\n";
$usage .= "\tOPTIONS:\n";
$usage .= "\t\t-T <n>       : minimum bit score to include is <n>\n";
$usage .= "\t\t-E <x>       : maximum E-value to include is <x>\n";
$usage .= "\t\t--cmscan     : tblout file was created by cmscan\n";
$usage .= "\t\t--source <s> : specify 'source' field should be <s> (e.g. tRNAscan-SE)\n";
$usage .= "\t\t--fmt2       : tblout file was created with cmscan --fmt 2 option\n";
$usage .= "\t\t--all        : output all info in 'attributes' column   [default: E-value]\n";
$usage .= "\t\t--none       : output no info in 'attributes' column    [default: E-value]\n";
$usage .= "\t\t--desc       : output desc field in 'attributes' column [default: E-value]\n";
$usage .= "\t\t--version <s>: append \"-<s>\" to 'source' column\n";
$usage .= "\t\t--extra <s>  : append \"<s>;\" to 'attributes' column\n";
$usage .= "\t\t--hidedesc   : do not includ \"desc\:\" prior to desc value in 'attributes' column\n";

my $do_minscore = 0;       # set to '1' if -T used
my $do_maxevalue = 0;      # set to '1' if -E used
my $minscore   = undef;    # defined if if -T used
my $maxevalue  = undef;    # defined if -E used
my $do_cmscan  = 0;        # set to '1' if --cmscan used
my $do_fmt2    = 0;        # set to '1' if --fmt
my $do_all_attributes = 0; # set to '1' if --all
my $do_no_attributes  = 0; # set to '1' if --none
my $do_de_attributes  = 0; # set to '1' if --desc
my $version = undef;       # defined if --version used
my $extra = undef;         # defined if --extra used
my $do_hidedesc = 0;       # set to 1 if --hidedesc used
my $opt_source = undef;    # set to <s> if --source <s> used

&GetOptions( "T=s"       => \$minscore,
             "E=s"       => \$maxevalue,
             "cmscan"    => \$do_cmscan,
             "source=s"  => \$opt_source,
             "fmt2"      => \$do_fmt2,
             "all"       => \$do_all_attributes,
             "none"      => \$do_no_attributes,
             "desc"      => \$do_de_attributes,
             "version=s" => \$version, 
             "extra=s"   => \$extra,
             "hidedesc"  => \$do_hidedesc);

if(scalar(@ARGV) != 1) { die $usage; }
my ($tblout_file) = @ARGV;

if(defined $minscore)  { $do_minscore  = 1; }
if(defined $maxevalue) { $do_maxevalue = 1; }
if($do_minscore && $do_maxevalue) { 
  die "ERROR, -T and -E cannot be used in combination. Pick one.";
}
if(($do_all_attributes) && ($do_no_attributes)) { 
  die "ERROR, --all and --none cannot be used in combination. Pick one.";
}
if(($do_all_attributes) && ($do_de_attributes)) { 
  die "ERROR, --all and --desc cannot be used in combination. Pick one.";
}
if(($do_no_attributes) && ($do_de_attributes)) { 
  die "ERROR, --none and --desc cannot be used in combination. Pick one.";
}
if(($do_fmt2) && (! $do_cmscan)) { 
  die "ERROR, --fmt2 only makes sense in combination with --cmscan"; 
}
if((defined $opt_source) && (defined $version)) { 
  die "ERROR, --source and --version are incompatible";
}

if(! -e $tblout_file) { die "ERROR tblout file $tblout_file does not exist"; }
if(! -s $tblout_file) { die "ERROR tblout file $tblout_file is empty"; }

my $source = ($do_cmscan) ? "cmscan" : "cmsearch";
if(defined $version)    { $source .= "-" . $version; }
if(defined $opt_source) { $source = $opt_source; }

open(IN, $tblout_file) || die "ERROR unable to open $tblout_file for reading"; 
my $line;
my $i;
while($line = <IN>) { 
  if($line !~ m/^\#/) { 
    chomp $line;
    my @el_A = split(/\s+/, $line);
    my $nfields = scalar(@el_A);
    if((! $do_fmt2) && ($nfields < 18)) { 
      die "ERROR expected at least 18 space delimited fields in tblout line (fmt 1, default) but got $nfields on line:\n$line\n"; 
    }
    if(($do_fmt2) && ($nfields < 27)) { 
      die "ERROR expected at least 27 space delimited fields in tblout line (fmt 2, --fmt2) but got $nfields on line:\n$line\n"; 
    }
    # ref Infernal 1.1.2 user guide, pages 59-61
    my $idx     = undef;
    my $seqname = undef;
    my $seqaccn = undef;
    my $mdlname = undef;
    my $mdlaccn = undef;
    my $clan    = undef;
    my $mdl     = undef;
    my $mdlfrom = undef;
    my $mdlto   = undef;
    my $seqfrom = undef;
    my $seqto   = undef;
    my $strand  = undef;
    my $trunc   = undef;
    my $pass    = undef;
    my $gc      = undef;
    my $bias    = undef;
    my $score   = undef;
    my $evalue  = undef;
    my $inc     = undef;
    my $olp     = undef;
    my $anyidx  = undef;
    my $anyfrct1= undef;
    my $anyfrct2= undef;
    my $winidx  = undef;
    my $winfrct1= undef;
    my $winfrct2= undef;
    my $desc    = undef;

    if($do_fmt2) { # 27 fields
      $idx     = $el_A[0];
      $seqname = ($do_cmscan) ? $el_A[3] : $el_A[1];
      $seqaccn = ($do_cmscan) ? $el_A[4] : $el_A[2];
      $mdlname = ($do_cmscan) ? $el_A[1] : $el_A[3];
      $mdlaccn = ($do_cmscan) ? $el_A[2] : $el_A[4];
      $clan    = $el_A[5];
      $mdl     = $el_A[6];
      $mdlfrom = $el_A[7];
      $mdlto   = $el_A[8];
      $seqfrom = $el_A[9];
      $seqto   = $el_A[10];
      $strand  = $el_A[11];
      $trunc   = $el_A[12];
      $pass    = $el_A[13];
      $gc      = $el_A[14];
      $bias    = $el_A[15];
      $score   = $el_A[16];
      $evalue  = $el_A[17];
      $inc     = $el_A[18];
      $olp     = $el_A[19];
      $anyidx  = $el_A[20];
      $anyfrct1= $el_A[21];
      $anyfrct2= $el_A[22];
      $winidx  = $el_A[23];
      $winfrct1= $el_A[24];
      $winfrct2= $el_A[25];
      $desc    = $el_A[26]; 
      for($i = 27; $i < $nfields; $i++) { $desc .= "_" . $el_A[$i]; }
    }
    else { # fmt 1, default
      $seqname = ($do_cmscan) ? $el_A[2] : $el_A[0];
      $seqaccn = ($do_cmscan) ? $el_A[3] : $el_A[1];
      $mdlname = ($do_cmscan) ? $el_A[0] : $el_A[2];
      $mdlaccn = ($do_cmscan) ? $el_A[1] : $el_A[3];
      $mdl     = $el_A[4];
      $mdlfrom = $el_A[5];
      $mdlto   = $el_A[6];
      $seqfrom = $el_A[7];
      $seqto   = $el_A[8];
      $strand  = $el_A[9];
      $trunc   = $el_A[10];
      $pass    = $el_A[11];
      $gc      = $el_A[12];
      $bias    = $el_A[13];
      $score   = $el_A[14];
      $evalue  = $el_A[15];
      $inc     = $el_A[16];
      $desc    = $el_A[17]; 
      for($i = 18; $i < $nfields; $i++) { $desc .= "_" . $el_A[$i]; }
    }
    # one sanity check, strand should make sense
    if(($strand ne "+") && ($strand ne "-")) { 
      if(($do_fmt2) && (($seqfrom eq "+") || ($seqfrom eq "-"))) { 
        die "ERROR problem parsing, you specified --fmt2 but tblout file appears to have NOT been created with --fmt 2, retry without --fmt2\nproblematic line:\n$line\n"; 
      }
      if((! $do_fmt2) && (($pass eq "+") || ($pass eq "-"))) { 
        die "ERROR problem parsing, you did not specify --fmt2 but tblout file appears to have been created with --fmt 2, retry with --fmt2\nproblematic line:\n$line\n"; 
      }
      die "ERROR unable to parse, problematic line:\n$line\n";
    }
    if(($do_minscore) && ($score < $minscore)) { 
      ; # skip
    }
    elsif(($do_maxevalue) && ($evalue > $maxevalue)) { 
      ; # skip
    }
    else { 
      my $attributes = "evalue=" . $evalue; # default to just evalue
      if($do_all_attributes) { 
        if($do_fmt2) { 
          $attributes .= sprintf(";idx=$idx;seqaccn=$seqaccn;mdlaccn=$mdlaccn;clan=$clan;mdl=$mdl;mdlfrom=$mdlfrom;mdlto=$mdlto;trunc=$trunc;pass=$pass;gc=$gc;bias=$bias;inc=$inc;olp=$olp;anyidx=$anyidx;anyfrct1=$anyfrct1;anyfrct2=$anyfrct2;winidx=$winidx;winfrct1=$winfrct1;winfrct2=$winfrct2;%s$desc", ($do_hidedesc ? "" : "desc="));
        }
        else { 
          $attributes .= sprintf(";seqaccn=$seqaccn;mdlaccn=$mdlaccn;mdl=$mdl;mdlfrom=$mdlfrom;mdlto=$mdlto;trunc=$trunc;pass=$pass;gc=$gc;bias=$bias;inc=$inc;%s$desc", ($do_hidedesc ? "" : "desc="));
        }
      }
      elsif($do_no_attributes) { 
        $attributes = "-";
      }
      elsif($do_de_attributes) { 
        $attributes = $desc;
      }
      if(defined $extra) { 
        if($attributes eq "-")       { $attributes = ""; }
        elsif($attributes !~ m/\;$/) { $attributes .= ";"; }
        $attributes .= $extra . ";";
      }
      printf("%s\t%s\t%s\t%d\t%d\t%.1f\t%s\t%s\t%s\n", 
             $seqname,                             # token 1: 'sequence' (sequence name)
             $source,                              # token 2: 'source'
             $mdlname,                             # token 3: 'feature' (model name) you may want to change this to 'ncRNA'
             ($strand eq "+") ? $seqfrom : $seqto, # token 4: 'start' in coordinate space [1..seqlen], must be <= 'end'
             ($strand eq "+") ? $seqto : $seqfrom, # token 5: 'end' in coordinate space [1..seqlen], must be >= 'start'
             $score,                               # token 6: 'score' bit score
             $strand,                              # token 7: 'strand' ('+' or '-')
             ".",                                  # token 8: 'phase' irrelevant for noncoding RNAs
             $attributes);                         # token 9: attributes, currently only E-value, unless --all, --none or --desc
    }
  }
}

附convert_tRNAScanSE_to_gff3.pl

#!/usr/bin/env perl

=head1 NAME
github:jorvis/biocode
tRNAScan_SE_to_gff3.pl - convert raw output of tRNAScan-SE to gff3
=head1 SYNOPSIS
USAGE: convert_tRNAScanSE_to_gff3.pl 
            --input=/path/to/some_file.out 
=head1 OPTIONS
B<--input,-i>
    The raw output from tRNAScan-SE:
    
    Sequence                        tRNA    Bounds  tRNA    Anti    Intron Bounds   Cove
    Name                    tRNA #  Begin   End     Type    Codon   Begin   End     Score
    --------                ------  ----    ------  ----    -----   -----   ----    ------
    tp.assembly.567468735.1         1       91820   91902   Tyr     GTA     91857   91866   66.58
    tp.assembly.567468735.1         2       171777  171849  Phe     GAA     0       0       70.28
    tp.assembly.567468735.1         3       172144  172215  His     GTG     0       0       64.04
    tp.assembly.567468735.1         4       852847  852919  Thr     AGT     0       0       75.69
    tp.assembly.567468735.1         5       877291  877362  Trp     CCA     0       0       68.97
    tp.assembly.567468735.1         6       1468229 1468300 Cys     GCA     0       0       72.10
    tp.assembly.567468735.1         7       2507459 2507530 Pro     AGG     0       0       62.33
    tp.assembly.567468735.1         8       2507198 2507127 Pro     CGG     0       0       65.73
    tp.assembly.567468735.1         9       2506317 2506246 Pro     TGG     0       0       66.60
    tp.assembly.567468735.1         10      2463785 2463713 Lys     TTT     0       0       79.47
    tp.assembly.567468735.1         11      2191149 2191069 Leu     CAG     0       0       57.47
    tp.assembly.567468735.1         12      1633307 1633237 Gly     CCC     0       0       65.52
    tp.assembly.567468735.1         13      1255051 1254968 Leu     CAA     0       0       60.46
    tp.assembly.567468735.1         14      251108  251037  Asp     GTC     0       0       59.48
    tp.assembly.567468735.1         15      250520  250449  Asp     GTC     0       0       59.48
B<--log,-l> 
    Log file
B<--ignoreIntrons,-g>
    Flag to ignore introns. tRNAs split by introns are usually flagged by NCBI's table2asn as too short.
B<--help,-h>
    This help message
=head1  DESCRIPTION
File converter
=head1  INPUT
Input above.
=head1  OUTPUT
GFF3 to STDOUT
=head1  CONTACT
    Kyle Tretina
    kyletretina@gmail.com
=cut

use warnings;
use strict;
use Getopt::Long qw(:config no_ignore_case no_auto_abbrev pass_through);
use Pod::Usage;

my %options = ();
my $results = GetOptions (\%options, 
                          'input|i=s',
                          'log|l=s',
                          'ignoreIntrons|g',
                          'help|h') || pod2usage();

## display documentation
if( $options{'help'} ){
    pod2usage( {-exitval => 0, -verbose => 2, -output => \*STDERR} );
}

## make sure everything passed was peachy
&check_parameters(\%options);

## open the log if requested
my $logfh;
if (defined $options{log}) {
    open($logfh, ">$options{log}") || die "can't create log file: $!";
}

## set ignoreIntrons flag
my $ignoreIntrons = '';
if (defined $options{ignoreIntrons}) {
    $ignoreIntrons = 1;
}


## open the input file
my $ifh;
open($ifh, "<$options{input}") || die "can't open input file: $!";

# all output needs the gff header
print "##gff-version 3\n";

## globals
my $tRNAGenNum=1;
my $tRNAExonNum=1;

## parse the file
foreach my $line (<$ifh>){
    my @cols = split /[\t]/, $line;
    chomp @cols;
    my $contig = $cols[0];

    if ($contig =~ /^(.+?)\s+$/) {
        $contig = $1;
    }

    ## skip the header lines
    next if $contig eq 'Sequence' || $contig eq 'Name' || $contig eq '--------';
    
    my $start = $cols[2] + 0;
    my $stop = $cols[3] + 0;
    my $trna = $cols[4];
    my $score = $cols[8];
    my $pseudo = "";
    my $anticodon = $cols[5];
    my $intron_start = $cols[6]-1; ## offset intron boundaries
    my $intron_end = $cols[7]+1;

    $pseudo = ";pseudogene=unknown" if $cols[9] =~ m/pseudo/;
    # Suppressor (sup) tRNAs are not recognized by NCBI's table2asn
    $trna = "Xxx" if $trna =~ m/Sup/ or $trna =~ m/Undet/;
    
    ## Check if tRNAScan-SE predicted a possible intron (value in column 6 > 0)
    if ($cols[6] > 0 and not $ignoreIntrons) {
        if ($start > $stop){
            my $intron_end = $cols[6]+1; ## offset intron boundaries
            my $intron_start = $cols[7]-1;
            ($start, $stop) = ($stop, $start);
        }
        print "$contig\ttRNAScan-SE\tgene\t$start\t$stop\t$score\t-\t.\tID=tRNA-$trna$tRNAGenNum\_gene$pseudo\n";
        print "$contig\ttRNAScan-SE\ttRNA\t$start\t$intron_start\t$score\t-\t.\tID=tRNA-$trna$tRNAExonNum\_tRNA;Parent=tRNA-$trna$tRNAGenNum\_gene;product=tRNA-$trna;anticodon=$anticodon;\n";
        $tRNAExonNum++;
        print "$contig\ttRNAScan-SE\ttRNA\t$intron_end\t$stop\t$score\t-\t.\tID=tRNA-$trna$tRNAExonNum\_tRNA;Parent=tRNA-$trna$tRNAGenNum\_gene;product=tRNA-$trna;anticodon=$anticodon;\n";
        $tRNAGenNum++;
        $tRNAExonNum++;
    } else{
        ($start, $stop) = ($stop, $start) if ($start > $stop);
        print "$contig\ttRNAScan-SE\tgene\t$start\t$stop\t$score\t-\t.\tID=tRNA-$trna$tRNAGenNum\_gene$pseudo\n";
        print "$contig\ttRNAScan-SE\ttRNA\t$start\t$stop\t$score\t-\t.\tID=tRNA-$trna$tRNAExonNum\_tRNA;product=tRNA-$trna;anticodon=$anticodon;Parent=tRNA-$trna$tRNAGenNum\_gene\n";
        $tRNAGenNum++;
        $tRNAExonNum++;
    }
}

exit(0);


sub _log {
    my $msg = shift;
    print $logfh "$msg\n" if $logfh;
}

sub check_parameters {
    my $options = shift;
    ## make sure required arguments were passed
    my @required = qw( input );
    for my $option ( @required ) {
        unless  ( defined $$options{$option} ) {
            die "--$option is a required option";
        }
    }
    ## handle some defaults
    $options{optional_argument2}   = 'foo'  unless ($options{optional_argument2});
}

https://www.jianshu.com/p/8f4061c38508
http://blog.genesino.com/2017/06/Rfam/