psathyrella
commented
7 years ago why don't you use 'all_matches'? then you won't have to modify the header consistency enforcement stuff. Can turn to ordered dict if you want to preserve order
On Mon, Oct 17, 2016 at 6:18 AM, Erick Matsen notifications@github.com wrote:
We need max and min of the qrbounds for each matched gene, as well as the relative alignment of the read to each matched gene. According to @psathyrella https://github.com/psathyrella , the best strategy is to put it into the infoline data structure.
For a qinfo that looks like this:
{'first_match_qrbounds': (0, 296), 'glbounds': {'IGHD2-15_01': (3, 31), 'IGHD2-2_03': (0, 31), 'IGHJ6_02': (3, 62), 'IGHJ6_03': (3, 62), 'IGHV1-2_02': (0, 296)}, 'matches': {'d': [(140, 'IGHD2-15_01'), (99, 'IGHD2-2_03')], 'j': [(288, 'IGHJ6_02'), (260, 'IGHJ6_03')], 'v': [(1473, 'IGHV1-2_02')]},'name': 'V1-2_04-D2-15_01-J6_01_del3d_Tins296_331_vins28', 'new_indel': False, 'qrbounds': {'IGHD2-15_01': (302, 330), 'IGHD2-2_03': (299, 330), 'IGHJ6_02': (336, 395), 'IGHJ6_03': (336, 395), 'IGHV1-2*02': (0, 296)}, 'seq': 'CAGGTGCAGCTGGTGCAGTCTGGGGCTGAGGTGAAGAAGCCTGGGGCCTCAGTGAAGGTCTCCTGCAAGGCTTCTGGATACACCTTCACCGGCTACTATATGCACTGGGTGCGACAGGCCCCTGGACAAGGGCTTGAGTGGATGGGATGGATCAACCCTAACAGTGGTGGCACAAACTATGCACAGAAGTTTCAGGGCTGGGTCACCATGACCAGGGACACGTCCATCAGCACAGCCTACATGGAGCTGAGCAGGCTGAGATCTGACGACACGGCCGTGTATTACTGTGCGAGAGATTTTTTATATTGTAGTGGTGGTAGCTGCTACTCCGGGGGGACTACTACTACTACGGTATGGACGTCTGGGGGCAAGGGACCACGGTCACCGTCTCCTCAG'}
and an indelinfo like this:
{'V1-2_04-D2-15_01-J6_01_del3d_Tins296_331_vins28': { 'indels': [{'len': 6, 'pos': 27, 'seqstr': 'CCCCCC', 'type': 'insertion'}], 'reversed_seq': 'CAGGTGCAGCTGGTGCAGTCTGGGGCTGAGGTGAAGAAGCCTGGGGCCTCAGTGAAGGTCTCCTGCAAGGCTTCTGGATACACCTTCACCGGCTACTATATGCACTGGGTGCGACAGGCCCCTGGACAAGGGCTTGAGTGGATGGGATGGATCAACCCTAACAGTGGTGGCACAAACTATGCACAGAAGTTTCAGGGCTGGGTCACCATGACCAGGGACACGTCCATCAGCACAGCCTACATGGAGCTGAGCAGGCTGAGATCTGACGACACGGCCGTGTATTACTGTGCGAGAGATTTTTTATATTGTAGTGGTGGTAGCTGCTACTCCGGGGGGACTACTACTACTACGGTATGGACGTCTGGGGGCAAGGGACCACGGTCACCGTCTCCTCAG'}}
I propose augmenting the infoline with two new items,
'qr_relpos': {'IGHD2-15_01': 299, 'IGHD2-2_03': 299, 'IGHJ6_02': 333, 'IGHJ6_03': 333, 'IGHV1-2*02': 0}
and
'qr_boundsbounds': {'d': (299, 330), 'j': (336, 395), 'v': (0, 296)}
The qr_relpos is simply the left hand qrbound minus the left hand glbound, and the boundsbounds is just the max and the min of each of the qrbounds across the genes.
There is already this:
swfo['regional_bounds'][region] = tuple([rb + padleft for rb in swfo['regional_bounds'][region]]) # I kind of want to just use a list now, but a.t.m. don't much feel like changing it everywhere elsewhich I don't totally get, but which evaluated on the above gives
'regional_bounds': {'d': (302, 330), 'j': (336, 395), 'v': (0, 296)},
so not exactly what I seem to want.
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/psathyrella/partis/issues/218, or mute the thread https://github.com/notifications/unsubscribe-auth/AC0rc9teFvIaO6oNpmwD5s8UKsHRxMpBks5q03WYgaJpZM4KYnJi .
matsen
dunleavy005
We need max and min of the
qrboundsfor each matched gene, as well as the relative alignment of the read to each matched gene. According to @psathyrella , the best strategy is to put it into the infoline data structure.For a
qinfothat looks like this:and an indelinfo like this:
I propose augmenting the
infolinewith two new items,and
The
qr_relposis simply the left hand qrbound minus the left hand glbound, and the boundsbounds is just the max and the min of each of the qrbounds across the genes.There is already this:
which I don't totally get, but which evaluated on the above gives
so not exactly what I seem to want.