Zulko
commented
6 years ago Hey there,
I agree more examples would be a good thing. There are several possible answers to your questions.
First, keep in mind that the feature objects in the Biotranslator refer to Biopython Feature objects. So the Biopython docs will tell you everything about their structure.
Also, if a feature has qualifiers "color" or "label", these will be used by default by the BioPythonTranslator (unless you have a custom BiopythonTranslator where you overwrite this behavior). That means that instead of putting your logics in your custom translator, you can also pre-process your biopython record before you feed it to the translator:
seqrecord = snapgene_file_to_seqrecord(args.infile)#
for feature in seqrecord:
feature.qualifiers["color"] = 'blue'
# ... implement other rules
graphic_record =Translator().translate_record(args.seqfile)This being said, here are some ways of doing what you want from the Translator:
For the gene product, I would do it as follows:
def compute_feature_label(self, feature):
if feature.type == "CDS":
return feature.qualifiers["product"]
# or even more robust, will handle the case with 0 or 2+ products:
return " ".join(feature.qualifiers.get("product", [""]))For iterating through colors, i would do it as follows:
from itertools import cycle
color_iterator = cycle(['blue', 'red', 'green', 'purple'])
def compute_feature_color(self, feature):
if feature.type == "CDS":
return next(color_iterator)also have a look at matplotlib.colors for ways of generating color palettes.
jrjhealey
MeggyC
Hi Zulko,
Love the package! Exactly what I've been looking for.
I have a couple of queries about the best way to implement the following things:
Firstly, I'd like to maybe colour each CDS on an individual basis using Matplotlib's colourscales. How would you advise implementing this? I've been using the code below, so I'm thinking I could maybe iterate a list of colours and zip it to the feature list so that each feature is paired with an individual colour? Any better suggestions?
Secondly, is there some way to access additional options for labelling features? In the documented examples, you've shown how to forcibly renaming all the CDS to a specific string (
"CDS here"). I'd like to label some/all of my features with the/producttag from a Genbank, as it's more informative than the locus tags. Is there some way to access these in the.featuresobject? I can't see something that looks like it corresponds to this.More generally, could I put a request in for some more 'fully featured' examples in the documentation that could be deconstructed (if you get the time of course!) ? I'd like to learn how to use the package in much more depth for the future.
Many thanks!