PostgreSQL performance

gistfile1.txt

EXPLAIN ANALYZE
SELECT COUNT(*) FROM j_msg

Row	QUERY PLAN
1	Aggregate  (cost=1360105.49..1360105.50 rows=1 width=0) (actual time=140042.984..140042.985 rows=1 loops=1)
2	  ->  Seq Scan on j_msg  (cost=0.00..1318772.99 rows=16532999 width=0) (actual time=8.377..137051.294 rows=16532999 loops=1)
3	Total runtime: 140057.576 ms

====

EXPLAIN ANALYZE
SELECT
    COUNT(msgid) msg_cnt,
    SUM(CASE WHEN parentmsgid IS NULL THEN 1 ELSE 0 END) parent_cnt,
    SUM(CASE WHEN parentmsgid IS NULL THEN 0 ELSE 1 END) child_cnt
FROM j_msg
WHERE created_at > (EXTRACT(EPOCH FROM DATE '2012-07-01') * 1000)

Row	QUERY PLAN
1	Aggregate  (cost=1442770.49..1442770.50 rows=1 width=16) (actual time=190405.935..190405.935 rows=1 loops=1)
2	  ->  Seq Scan on j_msg  (cost=0.00..1401437.98 rows=5511000 width=16) (actual time=634.916..190379.961 rows=45578 loops=1)
3	        Filter: ((created_at)::double precision > 1341118800000::double precision)
4	Total runtime: 190406.077 ms

====

table_name	indej_name	column_name
j_msg	j_msg_pk	msgid
j_msg	jvmssg_cdate_idx	created_at
j_msg	jvmssg_cidctmd_idx	c_type
j_msg	jvmssg_cidctmd_idx	c_id
j_msg	jvmssg_cidctmd_idx	updated_at
j_msg	jvmssg_mdate_idx	updated_at
j_msg	jvmssg_mdvle_idx	m_value
j_msg	jvmssg_prntid_idx	parentmsgid
j_msg	jvmssg_thrd_idx	threadid
j_msg	jvmssg_usrid_idx	userid