annotate macro.xml @ 0:bf0ebc9921f2 draft default tip

planemo upload for repository https://github.com/ARTbio/tools-artbio/tree/master/tools/wisecondorx commit b391b0f3348bd86a5c276dc4d3ff9dc98890c115

<macros>
    <token name="@VERSION@">1.2.9</token>
    <token name="@WRAPPER_VERSION@">@VERSION@+galaxy0</token>
    <token name="@PROFILE@">23.0</token>
    <token name="@pipefail@"><![CDATA[set -o | grep -q pipefail && set -o pipefail;]]></token>

    <xml name="requirements">
        <requirements>
            <requirement type="package" version="@VERSION@">wisecondorx</requirement>
        </requirements>
    </xml>
<token name="@help@"><![CDATA[
**What it does**

WisecondorX, which uses a within-sample normalization technique, detects Copy
Number Variation from BAM input files.

It is important that **no** read quality filtering is executed prior to running
WisecondorX: this software requires low-quality reads to distinguish informative
bins from non-informative ones.

There are three main stages (converting, reference build and predicting) when
using WisecondorX:

**1. Convert .bam files** of aligned reads to .npz files (for both normal and
tumor samples) using the Galaxy tool **WisecondorX convert bam to npz**

**2. Buid a reference index** from .npz files from **normal** samples using the
Galaxy tool **WisecondorX build reference**.

.. class:: warningmark

Automated gender prediction, required to consistently analyze sex chromosomes,
is based on a Gaussian mixture model. If few samples (<20) are included during
reference creation, or not both male and female samples (for NIPT, this means
male and female feti) are represented, this process might not be accurate.
Therefore, alternatively, one can manually tweak the --yfrac parameter.

.. class:: warningmark

It is of paramount importance that the reference set consists of exclusively
negative (normal) control samples that originate from the same sequencer, mapper,
reference genome, type of material, ... etc, as the test samples. As a rule of
thumb, think of all laboratory and in silico steps: the more sources of bias that
can be omitted, the better.

Try to include at least 50 samples per reference. The more the better, yet, from
500 on it is unlikely to observe additional improvement concerning normalization.

**3. Predict Copy Number Variantions** from the reference index and tumor .npz cases
of interest using the Galaxy tool **WisecondorX predict CNVs**

]]></token>
</macros>