I looked at those 2015 LA Marathon results and compared Meza's splits to some runners who finished in similar times. Here are the 5k/10k/15k/20k/25k/30k/35k/40k splits of those runners:
30545 Francisco Quijada (2:53:17) - 20:35, 20:36, 19:47, 19:26, 19:35, 20:27, 21:58, 21:14
30087 Fernando Navarro (2:54:54) 19:51, 19:40, 19:33, 19:36, 20:32, 21:04, 22:28, 22:28
1091 Kevin Purcell (2:53:46) 22:14, 21:41, 20:49, 20:25, 20:11, 20:18, 20:19, 19:36
22137 Danny Connolly (also 2:53:46) 20:20, 20:27, 20:05, 19:56, 20:38, 21:08, 21:21, 21:04
and here are Frank Meza's splits:
1287 Frank Meza (2:52:47) 20:47, 19:03, 20:28, 25:01, 18:34, 19:22, 21:28, 18:30
Pretty damning, if you ask me. The other guys all ran plausible, normal marathons. Their trajectories are recognizable to anyone who has a decent amount of experience running or observing marathons: generally a cautious start, followed by a gradual slight speeding up, but nothing dramatic. A middle period of very consistent splits. A couple of them faded at the end, as you would expect. But none of their 5k splits vary by much more than a minute from the prior 5k split.
Then you have Meza, who after the 1st 5k, goes -1:44, +1:25, +4:33, -6:27, +0:48, +2:06, -2:58 for his subsequent 5k splits. He's all over the place! Not plausible, and would require more than a single bathroom break to account for. Also, some of his outright 5k splits (18:30, 18:34) would be shockingly fast by themselves. I'm not buying it. And I'm taking it for granted that if he cheated in that race, you might as well throw out any other fast results.