You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
<td>Assume R2 starts with probes this long, and trim R1 bases that overlap into this (Default is 0)</td>
1667
1668
</tr>
@@ -1736,6 +1737,13 @@ <h3>petrim<a class="headerlink" href="#petrim" title="Permalink to this headline
1736
1737
bases that overlap into the first 11 bases of the left arm. Similar
1737
1738
trimming is available for situations where probes are ligated to the
1738
1739
right arm by using <code class="docutils literal"><span class="pre">--right-probe-length</span></code>.</li>
1740
+
<li>Adjustment of mismatching read bases inside areas of overlap. Such
1741
+
mismatches indicate that one or other of the bases has been
1742
+
incorrectly sequenced. Alteration of these bases is selected by
1743
+
supplying the <code class="docutils literal"><span class="pre">--mismatch-adjustment</span></code> flag with a value of
1744
+
<code class="docutils literal"><span class="pre">zero-phred</span></code> to alter the phred quality score of both bases to zero,
1745
+
or <code class="docutils literal"><span class="pre">pick-best</span></code> to choose whichever base had the higher reported
1746
+
quality score.</li>
1739
1747
<li>Removal of overlap regions by trimming both arms back to a point where
1740
1748
no overlap is present. An equal number of bases are removed from each
1741
1749
arm. This trimming is enabled by specifying <code class="docutils literal"><span class="pre">--midpoint-trim</span></code> and
@@ -11060,9 +11068,9 @@ <h4>Benchmarking comparisons using ROC and precision/sensitivity curves<a class=
11060
11068
and <a class="reference external" href="https://en.wikipedia.org/wiki/Sensitivity_and_specificity">https://en.wikipedia.org/wiki/Sensitivity_and_specificity</a></p>
<h4>Benchmarking performance for SNPs vs indels<a class="headerlink" href="#benchmarking-performance-for-snps-vs-indels" title="Permalink to this headline">¶</a></h4>
11065
-
<p>A common desire is to perform analysis separately for SNPs vs
<h4>Benchmarking performance for SNPs versus indels<a class="headerlink" href="#benchmarking-performance-for-snps-versus-indels" title="Permalink to this headline">¶</a></h4>
11073
+
<p>A common desire is to perform analysis separately for SNPs versus
11066
11074
indels. However, it is important to note that due the representation
11067
11075
ambiguity problem, it is not always trivial to decide in a global sense
11068
11076
whether a variant is a SNP or an indel or other complex variant. A group
@@ -11098,7 +11106,7 @@ <h4>Benchmarking performance for SNPs vs indels<a class="headerlink" href="#benc
11098
11106
another), then at some level there is really a semantic difference
11099
11107
between a “baseline indel” and a “call-set indel” (or “variant-caller-A
11100
11108
indel” and “variant-caller-B indel”), so caution should be applied when
11101
-
making conclusions related to SNP vs indel accuracy.</p>
11109
+
making conclusions related to SNP versus indel accuracy.</p>
11102
11110
<p>In the <code class="docutils literal"><span class="pre">snp_roc.tsv</span></code> and <code class="docutils literal"><span class="pre">non_snp_roc.tsv</span></code> output files, <code class="docutils literal"><span class="pre">vcfeval</span></code>
11103
11111
notes the number of baseline and call variants of each variant
11104
11112
type. When considering benchmarking metrics in the absence of any
@@ -11201,12 +11209,13 @@ <h3>svdecompose<a class="headerlink" href="#svdecompose" title="Permalink to thi
11201
11209
</table>
11202
11210
<p><strong>Usage:</strong></p>
11203
11211
<p>The <code class="docutils literal"><span class="pre">svdecompose</span></code> command is applied to a VCF containing structural
11204
-
variants and converts deletion, inversion, and tandem duplications with
11205
-
<code class="docutils literal"><span class="pre">SVTYPE</span></code> of <code class="docutils literal"><span class="pre">DEL</span></code>, <code class="docutils literal"><span class="pre">INV</span></code>, and <code class="docutils literal"><span class="pre">DUP</span></code>, respectively, into
11206
-
corresponding breakend events with <code class="docutils literal"><span class="pre">SVTYPE=BND</span></code>. <code class="docutils literal"><span class="pre">svdecompose</span></code> will
11207
-
also decompose sequence-resolved insertions and deletions greater than
11208
-
<code class="docutils literal"><span class="pre">--min-indel-length</span></code> into breakend representation. Records of others
11209
-
types are passed through without modification.</p>
11212
+
variants and converts deletion, insertion, inversion, and tandem
11213
+
duplications with <code class="docutils literal"><span class="pre">SVTYPE</span></code> of <code class="docutils literal"><span class="pre">DEL</span></code>, <code class="docutils literal"><span class="pre">INS</span></code>, <code class="docutils literal"><span class="pre">INV</span></code>, and
11214
+
<code class="docutils literal"><span class="pre">DUP</span></code>, respectively, into corresponding breakend events with
11215
+
<code class="docutils literal"><span class="pre">SVTYPE=BND</span></code>. <code class="docutils literal"><span class="pre">svdecompose</span></code> will also decompose sequence-resolved
11216
+
insertions and deletions greater than <code class="docutils literal"><span class="pre">--min-indel-length</span></code> into
11217
+
breakend representation. Records of others types are passed through
11218
+
without modification.</p>
11210
11219
<p>This operation can be useful for the purposes of reducing output from
11211
11220
various structural variant callers to a common representation to
11212
11221
better facilitate comparison with the <code class="docutils literal"><span class="pre">bndeval</span></code> command.</p>
0 commit comments