CMSC498Y meets on Mondays and Wednesdays.
The course schedule will be updated here, along with links to materials and assignments.
Slides will be posted to ELMS before the lecture and then linked here after the lecture.
If lecture is recorded, the video will typically be available under the zoom tab on ELMS. The overall schedule will be similar to spring 2024, with some improvements to pacing and content. There will also be 2 in-class midterms this year instead of 1.
BSA stands for the Biological Sequence Analysis textbook by Durbin, Eddy, Krogh, and Mitchison.
DL stands for the Deep Learning textbook by Ian Goodfellow, Yuoshua Bengio, and Aaron Courville.
| Week | Day | Date | Module | Topic | Materials | Assigned Reading |
|---|---|---|---|---|---|---|
| 1 | Mon | Jan 27 | Welcome! |
Course Overview, Policies, and Background Biology
[1_course_overview_and_biology.pdf] |
None | |
| Wed | Jan 29 | A | Random Sequence Model |
Random Sequence Model and Statistics Review (e.g., models, data, likelihood, maximum likelihood estimation, KL divergence)
[2_random_sequence_model_and_stats_review.pdf] Assignment #1 released. |
BSA Sections 1.3, 11.1 (through multinomial), 11.2 (relative entropy only), 11.3 (ML only), 11.5 (ML only), [bonus_probability_review.pdf] (optional) | |
| 2 | Mon | Feb 3 | A | Markov Models |
Bayes Classifier, Markov Models, Hidden Markov Models (definition only)
[3_markov_models.pdf] |
BSA Sections 3.1, 3.2 (through formal definition) |
| Wed | Feb 5 | A | No class |
Free period to work on Assignment #1
Add/drop deadline in two days (Fri Feb 7) |
None | |
| 3 | Mon | Feb 10 | A | Decoding HMMs |
Viterbi algorithm, Posterior decoding (Forward algorithm, Backward algorithm)
[4_decoding_hmms.pdf] |
BSA Section 3.2 |
| Wed | Feb 12 | B | MSAs |
Multiple Sequence Alignment (MSA) definition, Sum-of-Pairs Error, Edit distance, Sum-of-Pairs alignment, Star Alignment Heuristic
[5_msa.pdf] |
None | |
| 4 | Mon | Feb 17 | B | Profile HMMs |
Unadjusted Sequence Profile, Profile HMMs, Supervised Training given MSA, Decoding with Viterbi algorithm
[6_profile_hmm.pdf] Assignment #2 released. |
BSA Chapter 5 (through 5.4) |
| Wed | Feb 19 | B | Training HMMs |
Supervised training, Psuedocounts, Unsupervised training, Viterbi training, Baum-Welch (Expectation-Maximization)
[7_hmm_training.pdf] |
BSA Section 3.3 | |
| 5 | Mon | Feb 24 | B | Lab |
Profile HMM Lab - Adversarial example of failure when using Viterbi to align query sequence to MSA through profile HMM
[8_profile_hmm_lab.pdf] [profile-hmm-lab.zip] |
Lab Manual PDF |
| Wed | Feb 26 | A/B | Midterm #1 Review |
Discussion of course material on midterm exam #1
IMPORTANT: No zoom recording available [midterm1_study_guide.pdf] [midterm1_equation_sheet.pdf] |
Study Guide PDF | |
| 6 | Mon | Mar 3 | A/B | Midterm Exam #1 | Midterm exam in-class covering modules A and B | |
| Wed | Mar 5 | C | RNA Secondary Structure |
RNA secondary structure definition, evolutionary constraints, information degeneracy, psuedoknots, input/output to prediction problem, accuracy calculations
[9_rna_secondary_structure.pdf] |
BSA Chapter 10 (through 10.2) | |
| 7 | Mon | Mar 10 | C | Grammars |
Grammars, Context Free Grammars, Moore vs. Mealy Machines, Stochastic Context Free Grammars (SCFG), Parsing Algorithms
[10_scfgs.pdf] |
BSA Chapter 9 (skip Section 9.4) |
| Wed | Mar 12 | C | Optimization |
Maximium Base Pairs (Nussinov's Algorithm) and Minimum Energy
[11_rna_opt.pdf] |
BSA Section 10.2 through first sub-section on Energy minimization (skip SCFG sub-section) | |
| 8 | Mon | Mar 17 | No class | Spring break | ||
| Wed | Mar 18 | No class | Spring break | |||
| 9 | Mon | Mar 24 | C | UFold |
UFold Input / Ouput, Feature Construction, Output Postprocessing
[12_ufold.pdf] Asssignment #3 released. |
(Optional) Ufold paper, CDPfold paper, E2Efold paper |
| Wed | Mar 26 | C | Lab |
UFold Lab, part 1: Software installation and basic usage
[ufold-lab] |
Lab Manual PDF | |
| 10 | Mon | Mar 31 | C | Neural Networks |
Feedforward neural networks, Cost functions (e.g. cross-entropy), Output units and activation functions (sigmoid, softmax, linear, ReLU), Optimization Methods
[13_feedforward_neural_networks.pdf] |
DL Chapter 6 |
| Wed | Apr 2 | C | UNet | UNet Encoder / Contraction Path (e.g., Convolution and Max Pool), UNet Decoder / Expansion Path (e.g., Convolution and Upsampling) | DL Chapter 9 | |
| 11 | Mon | Apr 7 | C | Lab |
UFold Lab, part 2: Model and feature ablation
|
Lab Manual PDF |
| Wed | Apr 9 | C | Midterm #2 Review |
Discussion of course material on midterm exam #2
IMPORTANT: No zoom recording available Drop with a W deadline in two days (Fri Apr 11) |
Study Guide PDF | |
| 12 | Mon | Apr 14 | C | Midterm Exam #2 | Midterm exam in-class covering module C | |
| Wed | Apr 16 | |||||
| 13 | Mon | Apr 21 | ||||
| Wed | Apr 23 | |||||
| 14 | Mon | Apr 28 | ||||
| Wed | Apr 30 | |||||
| 15 | Mon | May 5 | ||||
| Wed | May 7 | |||||
| 16 | Mon | May 12 | Final Review | TBD | Study Guide PDF | |
| Wed | May 4 | No class | Reading day | |||
| Mon | May 19 | Final exam 4-6pm |