Aggregation
DeepBlue can aggregate groups of regions into larger regions.
The aggregate command requires three parameters: query_data_id, query_regions_id, and the data column name that will be the aggregation pivot.
It is possible to retrieve the aggregation result with the aggregation metafield:
| Aggregation Command | Description |
|---|---|
| @AGG.MIN | Regions minimum value |
| @AGG.MAX | Regions maximum value |
| @AGG.MEDIAN | Regions median value |
| @AGG.MEAN | Regions mean |
| @AGG.VAR | Regions variance |
| @AGG.SD | Regions standard deviation |
| @AGG.COUNT | Regions count |
Please, remember that DeepBlue does not aggregate inter-region regions, only the regions that are fully overlapped by some query_region_id.
In the following example, we aggregate the retrieved data into tiling regions of length 100000. In the, end we remove the aggregated regions that do not contain any region:
import xmlrpclib
import time
url = "http://deepblue.mpi-inf.mpg.de/xmlrpc"
server = xmlrpclib.Server(url, allow_none=True)
user_key = "anonymous_key"
(status, experiments) = server.list_experiments("hg19", None,
"DNA Methylation", "GM19240", "",
"RRBS", "ENCODE", user_key)
experiments_name = [e[1] for e in experiments]
(status, data_id) = server.select_regions(experiments_name, None, None,
None, None, None,
"chr22", None, None, user_key)
(s, regions_id) = server.tiling_regions(100000, "hg19", "chr22", user_key)
(status, aggr_id) = server.aggregate(data_id, regions_id, "SCORE", user_key)
(status, aggr_filter) = server.filter_regions(aggr_id, "@AGG.COUNT",
">", "0", "number", user_key)
(status, request_id) = server.get_regions(aggr_filter,
"CHROMOSOME,START,END,@AGG.MIN,@AGG.MAX, @AGG.MEDIAN,@AGG.MEAN,@AGG.SD,@AGG.COUNT",
user_key)
# Wait for the server processing
(status, info) = server.info(request_id, user_key)
request_status = info[0]["state"]
while request_status != "done" and request_status != "failed":
time.sleep(1)
(status, info) = server.info(request_id, user_key)
request_status = info[0]["state"]
print request_status
(status, regions) = server.get_request_data(request_id, user_key)
print regions