genome-minimal-fail.gff

##gff-version 3
scaffold_28	prediction	gene	1	402	0	+	.	ID=545184;Name=545184
scaffold_28	prediction	gene	805	981	0	-	.	ID=93782;Name=93782
scaffold_28	prediction	gene	2030	2721	0	+	.	ID=545205;Name=545205
scaffold_28	prediction	gene	3273	3545	0	-	.	Name=YOL159C-A;Synteny=no_synteny;SystematicGeneName=YOL159C-A;ID=38792
scaffold_28	prediction	gene	5318	5833	0	-	.	Name=YOL159C;Synteny=no_synteny;SystematicGeneName=YOL159C;ID=38793
scaffold_28	prediction	gene	6780	8600	0	-	.	Name=ENB1;Synteny=no_synteny;SystematicGeneName=YOL158C;StandardGeneName=ENB1;ID=38794
scaffold_28	prediction	gene	9698	11467	0	-	.	Name=IMA4;Synteny=no_synteny;SystematicGeneName=YJL221C;StandardGeneName=IMA4;ID=38795

minimal-fail.py

import gffutils
import os

def updateGenome():
    genomeGFF="genome-minimal-fail.gff"
    genomeGFFdb = genomeGFF.replace(".gff", ".db")

    #create database for genome
    if os.path.isfile(genomeGFFdb):
        db = gffutils.FeatureDB(genomeGFFdb)
    else:
        db = gffutils.create_db(genomeGFF, dbfn=genomeGFFdb, force=True, merge_strategy="merge")

    #shift coordinates of features that follow the original integration site by deletion size
    downstreamGenomeFeatures = db.region('scaffold_28:6001-11467', completely_within=True)

    for feature in downstreamGenomeFeatures:
        print(feature["Name"][0])
        feature.start += 10000
        feature.end += 10000
        print(feature.end)

    db.update(downstreamGenomeFeatures, merge_strategy="replace")

    return

updateGenome()

I can't figure out how to submit a web-based pull request on a gist, so I'll just describe here:

downstreamGenomeFeatures on line 15 is a generator. Like all generators (and just like reading a file object), iterating through it will consume it. You can't rewind. So on line 17 when you iterate over it, it consumes everything in the generator. Then when you call db.update and give it downstreamGenomeFeatures, by that time it's empty. So you get the error, "ValueError: No lines parsed -- was an empty file provided?" because no features were provided.

One solution is to make your own generator that consumes downstreamGenomeFeatures but yields the modified feature. Then give that generator to db.update. Like this:

    def gen():
        for feature in downstreamGenomeFeatures:
            print(feature["Name"][0])
            feature.start += 10000
            feature.end += 10000
            print(feature.end)
        yield feature

    db.update(gen(), merge_strategy="replace")

Thanks a lot. Could you please check my new example, though, at https://gist.github.com/rmzelle/22fc9c537ddf8f40de4b

I'm trying to shift the last two features of the GFF file downstream by 10000 bp. While the generator shows that the coordinates change, somehow the changes don't make it into the database after running db.update().