Maximum Likelihood - Revision history

Gorm at 13:14, 11 March 2026

2026-03-11T13:14:25Z

← Older revision		Revision as of 15:14, 11 March 2026
Line 73:		Line 73:
	----		----

	'''Question 4''' Use chatGPT (or some other large language model/AI) to help you figure out how to use the software iqtree3 (which is installed on your computer) to do the same analysis, i.e., to construct a maximum likelihood tree using the K2P + gamma model. Compare the output from ~~the two programs~~ to see if they agree.		'''Question 4''' Use chatGPT (or some other large language model/AI) to help you figure out how to use the software iqtree3 (which is installed on your computer) to do the same analysis, i.e., to construct a maximum likelihood tree using the K2P + gamma model. Show your prompt and the response.

			----

			'''Question 5''' Now actually run iqtree3 in the way you just figured out. Compare the output from PAUP and iqtree3 to see if they agree.

Gorm at 12:57, 11 March 2026

2026-03-11T12:57:08Z

← Older revision		Revision as of 14:57, 11 March 2026
Line 60:		Line 60:
	----		----

	'''Question 2''' Is the Neanderthal sequence placed inside or outside the clade of human sequences?		'''Question 2'''

			In PAUP issue this command to get estimates of model parameters (and the log likelihood of the fitted model):
			lscores

			'''Question: ''' What are the maximum likelihood estimates of the two model parameters (transition transversion ratio parameter kappa, and gamma shape parameter alpha)?

			----

			'''Question 3''' Is the Neanderthal sequence placed inside or outside the clade of human sequences?

	----		----

	'''Question 3''' Use chatGPT (or some other large language model/AI) to help you figure out how to use the software iqtree3 (which is installed on your computer) to do the same analysis, i.e., to construct a maximum likelihood tree using the K2P + gamma model. Compare the output from the two programs to see if they agree.		'''Question 4''' Use chatGPT (or some other large language model/AI) to help you figure out how to use the software iqtree3 (which is installed on your computer) to do the same analysis, i.e., to construct a maximum likelihood tree using the K2P + gamma model. Compare the output from the two programs to see if they agree.

Gorm at 12:50, 11 March 2026

2026-03-11T12:50:05Z

← Older revision		Revision as of 14:50, 11 March 2026
Line 64:		Line 64:
	----		----

	'''Question 3''' Use chatGPT (or some other large language model/AI) to help you figure out how to use the software iqtree3 (which is installed on your computer) to do the same analysis, i.e., to construct a maximum likelihood tree using the K2P + gamma model		'''Question 3''' Use chatGPT (or some other large language model/AI) to help you figure out how to use the software iqtree3 (which is installed on your computer) to do the same analysis, i.e., to construct a maximum likelihood tree using the K2P + gamma model. Compare the output from the two programs to see if they agree.

Gorm at 12:49, 11 March 2026

2026-03-11T12:49:29Z

← Older revision		Revision as of 14:49, 11 March 2026
Line 61:		Line 61:

	'''Question 2''' Is the Neanderthal sequence placed inside or outside the clade of human sequences?		'''Question 2''' Is the Neanderthal sequence placed inside or outside the clade of human sequences?

			----

			'''Question 3''' Use chatGPT (or some other large language model/AI) to help you figure out how to use the software iqtree3 (which is installed on your computer) to do the same analysis, i.e., to construct a maximum likelihood tree using the K2P + gamma model

Gorm: /* Getting started */

2025-03-11T19:35:48Z

Getting started

← Older revision		Revision as of 21:35, 11 March 2025
Line 52:		Line 52:
	'''Start heuristic search of tree space using Nearest Neighbour Interchange (NNI)'''		'''Start heuristic search of tree space using Nearest Neighbour Interchange (NNI)'''
	hsearch swap=nni		hsearch swap=nni
			describetrees all/plot=phylogram labelnode=no
	: This step may take a little while to finish (depending on your computer). For large datasets you sometimes have to wait hours or even days for a maximum likelihood analysis to finish. To save time we have here used swap=nni, which is less thorough than spr and ~~tbs~~, and typically faster. If you wanted to be more certain of finding the best tree you should use swap=tbr.		: This step may take a little while to finish (depending on your computer). For large datasets you sometimes have to wait hours or even days for a maximum likelihood analysis to finish. To save time we have here used swap=nni, which is less thorough than swap=spr and swap=tbr, and typically faster. If you wanted to be more certain of finding the best tree you should use swap=tbr.
	: The score that PAUP lists for maximum likelihood analysis is the ''negative'' log of the likelihood, -ln(L). (Recall that since likelihoods are numbers between 0 and 1, log-likelihoods will be negative numbers, and therefore negative log-likelihoods will be positive numbers. Perhaps a bit confusing that PAUP doesn't simply list the ln(L) ). As the likelihood increases, this number will decrease.		: The score that PAUP lists for maximum likelihood analysis is the ''negative'' log of the likelihood, -ln(L). (Recall that since likelihoods are numbers between 0 and 1, log-likelihoods will be negative numbers, and therefore negative log-likelihoods will be positive numbers. Perhaps a bit confusing that PAUP doesn't simply list the ln(L) ). As the likelihood increases, this number will decrease.

Gorm: /* Getting started */

2025-03-11T19:34:41Z

Getting started

← Older revision		Revision as of 21:34, 11 March 2025
Line 26:		Line 26:

	'''Remove subset of sequences to reduce computational burden:'''		'''Remove subset of sequences to reduce computational burden:'''
	delete 5-34		delete 5-40
	: This command removes 30 human sequences (sequence number 5 to sequence number 34) from the data set. We do this in order to reduce the time needed to finish the analysis. In the remaining data set we now have 23 human sequences, one chimpanzee, one bonobo, and one Neanderthal.		: This command removes 36 human sequences (sequence number 5 to sequence number 40) from the data set. We do this in order to reduce the time needed to finish the analysis. In the remaining data set we now have 17 human sequences, one chimpanzee, one bonobo, and one Neanderthal.

	'''Specify substitution model'''		'''Specify substitution model'''
Line 50:		Line 50:
	: The chimpanzee and the bonobo form the outgroup		: The chimpanzee and the bonobo form the outgroup

	'''Start heuristic search of tree space using ~~Tree Bisection and Reconnection~~ (~~TBR~~)'''		'''Start heuristic search of tree space using Nearest Neighbour Interchange (NNI)'''
	hsearch swap=~~tbr~~		hsearch swap=nni
	: This step may take a little while to finish (depending on your computer). For large datasets you sometimes have to wait hours or even days for a maximum likelihood analysis to finish. The score that PAUP lists for maximum likelihood analysis is the ''negative'' log of the likelihood, -ln(L). (Recall that since likelihoods are numbers between 0 and 1, log-likelihoods will be negative numbers, and therefore negative log-likelihoods will be positive numbers. Perhaps a bit confusing that PAUP doesn't simply list the ln(L) ). As the likelihood increases, this number will decrease.
			: This step may take a little while to finish (depending on your computer). For large datasets you sometimes have to wait hours or even days for a maximum likelihood analysis to finish. To save time we have here used swap=nni, which is less thorough than spr and tbs, and typically faster. If you wanted to be more certain of finding the best tree you should use swap=tbr.
			: The score that PAUP lists for maximum likelihood analysis is the ''negative'' log of the likelihood, -ln(L). (Recall that since likelihoods are numbers between 0 and 1, log-likelihoods will be negative numbers, and therefore negative log-likelihoods will be positive numbers. Perhaps a bit confusing that PAUP doesn't simply list the ln(L) ). As the likelihood increases, this number will decrease.

	'''Question: '''What is the negative log-likelihood, -ln(L), for the best tree found using NNI?		'''Question: '''What is the negative log-likelihood, -ln(L), for the best tree found using NNI?

Gorm: /* Getting started */

2025-03-11T19:27:02Z

Getting started

← Older revision		Revision as of 21:27, 11 March 2025
Line 42:		Line 42:

	: Finally, we specify that we want to use a model where substitution rates at different sites follow a gamma distribution (rates=gamma), and that we want the shape of this distribution to also be estimated from the data (shape=estimate).		: Finally, we specify that we want to use a model where substitution rates at different sites follow a gamma distribution (rates=gamma), and that we want the shape of this distribution to also be estimated from the data (shape=estimate).

			: '''Note:''' If we had omitted "tratio=estimate" and "shape=estimate", then PAUP would have used fixed, default values for these two parameters, rather than optimising them by maximum likelihood.

	'''Specify outgroup and rooting options'''		'''Specify outgroup and rooting options'''

Gorm: /* Getting started */

2025-03-11T19:23:51Z

Getting started

← Older revision		Revision as of 21:23, 11 March 2025
Line 48:		Line 48:
	: The chimpanzee and the bonobo form the outgroup		: The chimpanzee and the bonobo form the outgroup

	'''Start heuristic search of tree space using ~~nearest neighbor interchange~~ (~~NNI~~)'''		'''Start heuristic search of tree space using Tree Bisection and Reconnection (TBR)'''
	hsearch swap=~~nni~~		hsearch swap=tbr
	: This step may take a little while to finish (depending on your computer). For large datasets you sometimes have to wait hours or even days for a maximum likelihood analysis to finish. The score that PAUP lists for maximum likelihood analysis is the ''negative'' log of the likelihood, -ln(L). (Recall that since likelihoods are numbers between 0 and 1, log-likelihoods will be negative numbers, and therefore negative log-likelihoods will be positive numbers. Perhaps a bit confusing that PAUP doesn't simply list the ln(L) ). As the likelihood increases, this number will decrease.		: This step may take a little while to finish (depending on your computer). For large datasets you sometimes have to wait hours or even days for a maximum likelihood analysis to finish. The score that PAUP lists for maximum likelihood analysis is the ''negative'' log of the likelihood, -ln(L). (Recall that since likelihoods are numbers between 0 and 1, log-likelihoods will be negative numbers, and therefore negative log-likelihoods will be positive numbers. Perhaps a bit confusing that PAUP doesn't simply list the ln(L) ). As the likelihood increases, this number will decrease.

Gorm: /* Getting started */

2025-03-11T19:21:15Z

Getting started

← Older revision		Revision as of 21:21, 11 March 2025
Line 26:		Line 26:

	'''Remove subset of sequences to reduce computational burden:'''		'''Remove subset of sequences to reduce computational burden:'''
	delete 5-40		delete 5-34
	: This command removes 36 human sequences (sequence number 5 to sequence number 40) from the data set. We do this in order to reduce the time needed to finish the analysis. In the remaining data set we now have 17 human sequences, one chimpanzee, one bonobo, and one Neanderthal.		: This command removes 30 human sequences (sequence number 5 to sequence number 34) from the data set. We do this in order to reduce the time needed to finish the analysis. In the remaining data set we now have 23 human sequences, one chimpanzee, one bonobo, and one Neanderthal.

	'''Specify substitution model'''		'''Specify substitution model'''

Gorm: /* Getting started */

2024-04-24T09:03:46Z

Getting started

← Older revision		Revision as of 11:03, 24 April 2024
Line 11:		Line 11:

	'''Construct working directory, copy files'''		'''Construct working directory, copy files'''
	In a terminal window enter (remember to replace /path/to/molevol with your actual path to the molevol folder):
			In a terminal window enter the following commands (remember to replace /path/to/molevol with your actual path to the molevol folder):
	cd /path/to/molevol		cd /path/to/molevol
	mkdir likelihood		mkdir likelihood