Motifs derived from filter-activating sequences

Motifs from filter weights

In [1]:
import os
import sys
sys.path.append(os.path.abspath("/users/amtseng/tfmodisco/src/"))
from motif.read_motifs import pfm_info_content, pfm_to_pwm
from util import figure_to_vdom_image
import plot.viz_sequence as viz_sequence
import h5py
import numpy as np
import pyfaidx
import matplotlib.pyplot as plt
import vdom.helpers as vdomh
from IPython.display import display
import tqdm
tqdm.tqdm_notebook()

/users/amtseng/miniconda3/envs/tfmodisco-mini/lib/python3.7/site-packages/ipykernel_launcher.py:14: TqdmDeprecationWarning: This function will be removed in tqdm==5.0.0
Please use `tqdm.notebook.tqdm` instead of `tqdm.tqdm_notebook`
Out[1]:
0it [00:00, ?it/s]

Define constants and paths

In [2]:
# Define parameters/fetch arguments
filter_activations_path = os.environ["TFM_FILTER_ACTIVATIONS"]
filter_weights_path = os.environ["TFM_FILTER_WEIGHTS"]

if "TFM_MOTIF_CACHE" in os.environ:
    activation_motifs_cache_dir = os.environ["TFM_MOTIF_CACHE"]
else:
    activation_motifs_cache_dir = None

print("Path to filter activations: %s" % filter_activations_path)
print("Path to filter weights: %s" % filter_weights_path)
print("Saved activation-derived motifs cache: %s" % activation_motifs_cache_dir)

Path to filter activations: /users/amtseng/tfmodisco/results/filter_activations/singletask_profile_finetune/NR3C1-reddytime_singletask_profile_finetune_task5_fold5_filter_activations.h5
Path to filter weights: /users/amtseng/tfmodisco/results/filter_weights/singletask_profile_finetune/NR3C1-reddytime_singletask_profile_finetune_task5_fold5_filter_weights.npy
Saved activation-derived motifs cache: /users/amtseng/tfmodisco/results/reports/filter_derived_motifs/cache/NR3C1-reddytime_singletask_profile_finetune_task5_fold5_filter
In [3]:
# Constants/paths
input_length = 2114
filter_width = 21
reference_genome_path = "/users/amtseng/genomes/hg38.fasta"
In [4]:
if activation_motifs_cache_dir:
    os.makedirs(activation_motifs_cache_dir, exist_ok=True)

Helper functions

For extracting motifs

In [5]:
def dna_to_one_hot(seqs):
    """
    Converts a list of DNA ("ACGT") sequences to one-hot encodings, where the
    position of 1s is ordered alphabetically by "ACGT". `seqs` must be a list
    of N strings, where every string is the same length L. Returns an N x L x 4
    NumPy array of one-hot encodings, in the same order as the input sequences.
    All bases will be converted to upper-case prior to performing the encoding.
    Any bases that are not "ACGT" will be given an encoding of all 0s.
    """
    seq_len = len(seqs[0])
    assert np.all(np.array([len(s) for s in seqs]) == seq_len)

    # Join all sequences together into one long string, all uppercase
    seq_concat = "".join(seqs).upper()

    one_hot_map = np.identity(5)[:, :-1]

    # Convert string into array of ASCII character codes;
    base_vals = np.frombuffer(bytearray(seq_concat, "utf8"), dtype=np.int8)

    # Anything that's not an A, C, G, or T gets assigned a higher code
    base_vals[~np.isin(base_vals, np.array([65, 67, 71, 84]))] = 85

    # Convert the codes into indices in [0, 4], in ascending order by code
    _, base_inds = np.unique(base_vals, return_inverse=True)

    # Get the one-hot encoding for those indices, and reshape back to separate
    return one_hot_map[base_inds].reshape((len(seqs), seq_len, 4))
In [6]:
def extract_filter_activation_motifs(filter_activations_path, reference_genome_path):
    """
    Extracts the motifs that correspond to each filter. Returns an
    F x W x 4 array, where F is the number of filters and W is the width
    of each filter. The order of filters matches those in the saved HDF5/model.
    """
    reader = h5py.File(filter_activations_path, "r")
    activations_reader = reader["activations"]
    num_coords, two, num_windows, num_filters = activations_reader.shape
    
    assert two == 2
    assert num_windows == input_length - filter_width + 1
    
    print("Importing coordinates...")
    coords = np.empty((num_coords, 3), dtype=object)
    coords[:, 0] = reader["coords"]["coords_chrom"][:].astype(str)
    coords[:, 1] = reader["coords"]["coords_start"][:]
    coords[:, 2] = reader["coords"]["coords_end"][:]
    
    print("Fetching one-hot sequences...")
    genome_reader = pyfaidx.Fasta(reference_genome_path)
    one_hot_seqs = np.empty((num_coords, input_length, 4))
    batch_size = 128
    num_batches = int(np.ceil(num_coords / batch_size))
    for i in tqdm.notebook.trange(num_batches):
        batch_slice = slice(i * batch_size, (i + 1) * batch_size)
        one_hot_seqs[batch_slice] = dna_to_one_hot([
            genome_reader[chrom][start:end].seq for chrom, start, end in coords[batch_slice]
        ])
    
    pfms = np.empty((num_filters, filter_width, 4))
    for filter_index in range(num_filters):
        print("Extracting motif for filter %d..." % filter_index)
    
        print("\tComputing maximum activation...")
        acts = activations_reader[:, :, :, filter_index]
        max_act = np.max(acts)
        
        inds = np.where(acts >= 0.5 * max_act)
        
        windows, num_windows = np.zeros((filter_width, 4)), 0
        for coord_index, strand_index, pos_index in tqdm.notebook.tqdm(
            zip(*inds), total=len(inds[0]), desc="Extracting windows..."
        ):
            if strand_index == 0:
                window = one_hot_seqs[coord_index, pos_index : pos_index + filter_width]
            else:
                # Reverse complement; the positions are flipped
                window = np.flip(
                    one_hot_seqs[coord_index, input_length - filter_width - pos_index : input_length - pos_index],
                    axis=(0, 1)
                )
            windows = windows + window
            num_windows += 1
        
        pfms[filter_index] = windows / num_windows
    
    return pfms
In [7]:
def compute_filter_influence(filter_activations_path):
    """
    Extracts the influence of each filter by computing the difference
    in cross entropy when each filter is nullified.
    Returns an F-array, where F is the number of filters, containing the
    change in average cross entropy (after nullification - before
    nullification). The order of filters matches those in the saved
    HDF5/model.
    """
    reader = h5py.File(filter_activations_path, "r")
    print("Reading in cross entropies...")
    before_null_cross_ents = reader["predictions"]["cross_ents"][:]
    after_null_cross_ents = reader["nullified_predictions"]["cross_ents"][:]
    
    before_null = np.nanmean(before_null_cross_ents)
    
    num_filters = after_null_cross_ents.shape[1]
    
    influences = []
    for filter_index in tqdm.notebook.trange(num_filters):
        after_null = np.nanmean(after_null_cross_ents[:, filter_index])
        influences.append(after_null - before_null)
        
    return np.array(influences)
In [8]:
def save_activation_motifs(filter_pfms, filter_influences, path):
    """
    Saves the filter-activation-derived PFMs and influence values.
    """
    with h5py.File(path, "w") as f:
        f.create_dataset("pfms", data=filter_pfms, compression="gzip")
        f.create_dataset("influences", data=filter_influences, compression="gzip")
In [9]:
def load_activation_motifs(path):
    """
    Loads the filter-activation-derived PFMs and influence values.
    """
    with h5py.File(path, "r") as f:
        return f["pfms"][:], f["influences"][:]

Extract motifs from filter activations

Extract the motifs derived from each filter, ranked by filter influence.

Deriving a filter's motif:

  1. Identify the top 10000 most well-predicted input sequences, ranked by cross entropy
  2. For each window in each of these sequences, compute the filter activation for each 1st-layer filter
  3. A filter's motif is the aggregation of sequence windows which activate that filter to at least half its maximum activation (over the top 10000 most well-predicted inputs)

Deriving a filter's influence:

  1. Identify the top 10000 most well-predicted input sequences, ranked by cross entropy
  2. Nullify each filter by setting it to the average activation over these 10000 most well-predicted inputs
  3. A filter's influence is the average change in cross entropy before and after nullification
In [10]:
compute_motifs = True
if activation_motifs_cache_dir:
    # Import if it exists
    cache_path = os.path.join(activation_motifs_cache_dir, "filter_activation_motifs.h5")
    if os.path.exists(cache_path) and os.stat(cache_path).st_size:
        filter_pfms, filter_influences = load_activation_motifs(cache_path)
        compute_motifs = False

if compute_motifs:
    # Extract PFMs of highly-activating sequences
    filter_pfms = extract_filter_activation_motifs(filter_activations_path, reference_genome_path)

    # Compute influence of each filter
    filter_influences = compute_filter_influence(filter_activations_path)

    if activation_motifs_cache_dir:
        save_activation_motifs(filter_pfms, filter_influences, cache_path)

Importing coordinates...
Fetching one-hot sequences...

Extracting motif for filter 0...
	Computing maximum activation...

Extracting motif for filter 1...
	Computing maximum activation...

Extracting motif for filter 2...
	Computing maximum activation...

Extracting motif for filter 3...
	Computing maximum activation...

Extracting motif for filter 4...
	Computing maximum activation...

Extracting motif for filter 5...
	Computing maximum activation...

Extracting motif for filter 30...
	Computing maximum activation...

Extract motifs from filter weights

In [11]:
# Import the filter weights themselves
filter_weights = np.load(filter_weights_path)
assert len(filter_weights.shape) == 3
assert filter_weights.shape[:2] == (filter_width, 4)
filter_weights = np.transpose(filter_weights, axes=(2, 0, 1))  # Shape: F x W x 4

Motifs derived from filter-activating sequences

For each filter, its motif is constructed by averaging all of the sequences that activate it at least to half of its maximal activation. We show the PWMs. The filters are ranked by influence (i.e. the average difference in prediction cross entropy when the filter is nullified--that is, replaced with its average activation).

In [12]:
colgroup = vdomh.colgroup(
    vdomh.col(style={"width": "5%"}),
    vdomh.col(style={"width": "5%"}),
    vdomh.col(style={"width": "5%"}),
    vdomh.col(style={"width": "85%"})
)
header = vdomh.thead(
    vdomh.tr(
        vdomh.th("Rank", style={"text-align": "center"}),
        vdomh.th("Filter index", style={"text-align": "center"}),
        vdomh.th("Influence", style={"text-align": "center"}),
        vdomh.th("PWM", style={"text-align": "center"})
    )
)

body = []
for i, filter_index in enumerate(np.flip(np.argsort(filter_influences))):
    pwm = pfm_to_pwm(filter_pfms[filter_index])
    if np.sum(pwm[:, [0, 2]]) < 0.5 * np.sum(pwm):
        # Flip to purine-rich version
        pwm = np.flip(pwm, axis=(0, 1))
    fig = viz_sequence.plot_weights(pwm, figsize=(20, 4), return_fig=True)
    fig.tight_layout()
    
    body.append(
        vdomh.tr(
            vdomh.td(str(i + 1)),
            vdomh.td(str(filter_index)),
            vdomh.td("%.3f" % filter_influences[filter_index]),
            vdomh.td(figure_to_vdom_image(fig))
        )
    )
    
    if activation_motifs_cache_dir:
        # Save motif PWM
        fig.savefig(os.path.join(activation_motifs_cache_dir, "filter_activation_motif_%d.png" % filter_index))

display(vdomh.table(colgroup, header, vdomh.tbody(*body)))
plt.close("all")

/users/amtseng/tfmodisco/src/plot/viz_sequence.py:152: RuntimeWarning: More than 20 figures have been opened. Figures created through the pyplot interface (`matplotlib.pyplot.figure`) are retained until explicitly closed and may consume too much memory. (To control this warning, see the rcParam `figure.max_open_warning`).
  fig = plt.figure(figsize=figsize)
RankFilter indexInfluencePWM
1160.055
260.042
3340.015
4180.012
5110.011
6380.010
7240.009
8580.008
9140.006
10170.006
11540.004
12200.004
13130.002
14570.001
15270.001
16330.001
17530.001
18450.000
1900.000
20350.000
21370.000
2240.000
2330.000
24280.000
25310.000
26490.000
27400.000
28430.000
299-0.000
3051-0.000
3115-0.000
3252-0.000
3355-0.000
3412-0.000
3556-0.000
3610-0.000
3739-0.000
3850-0.000
397-0.000
4059-0.000
415-0.000
4260-0.000
4361-0.000
442-0.000
451-0.000
468-0.000
4719-0.000
4848-0.000
4929-0.000
5036-0.000
5141-0.000
5242-0.000
5332-0.000
5462-0.000
5530-0.000
5644-0.000
5721-0.000
5846-0.000
5926-0.000
6025-0.000
6147-0.000
6223-0.000
6322-0.000
6463-0.000

Motifs derived from filter weights

For each filter, we show its corresponding motif simply as the mean-normalized multiplicative weights in the filter. For consistency, we rank the filters by influence (as above).

In [13]:
colgroup = vdomh.colgroup(
    vdomh.col(style={"width": "5%"}),
    vdomh.col(style={"width": "5%"}),
    vdomh.col(style={"width": "5%"}),
    vdomh.col(style={"width": "85%"})
)
header = vdomh.thead(
    vdomh.tr(
        vdomh.th("Rank", style={"text-align": "center"}),
        vdomh.th("Filter index", style={"text-align": "center"}),
        vdomh.th("Influence", style={"text-align": "center"}),
        vdomh.th("Mean-normalized filter weights", style={"text-align": "center"})
    )
)

body = []
for i, filter_index in enumerate(np.flip(np.argsort(filter_influences))):
    weights = filter_weights[filter_index]
    weights = weights - np.mean(weights, axis=1, keepdims=True)
    if np.sum(weights[:, [0, 2]]) < 0.5 * np.sum(weights):
        # Flip to purine-rich version
        weights = np.flip(weights, axis=(0, 1))
    fig = viz_sequence.plot_weights(weights, figsize=(20, 4), return_fig=True)
    fig.tight_layout()
    
    body.append(
        vdomh.tr(
            vdomh.td(str(i + 1)),
            vdomh.td(str(filter_index)),
            vdomh.td("%.3f" % filter_influences[filter_index]),
            vdomh.td(figure_to_vdom_image(fig))
        )
    )
    
    if activation_motifs_cache_dir:
        # Save motif PWM
        fig.savefig(os.path.join(activation_motifs_cache_dir, "filter_weight_motif_%d.png" % filter_index))

display(vdomh.table(colgroup, header, vdomh.tbody(*body)))
plt.close("all")
RankFilter indexInfluenceMean-normalized filter weights
1160.055
260.042
3340.015
4180.012
5110.011
6380.010
7240.009
8580.008
9140.006
10170.006
11540.004
12200.004
13130.002
14570.001
15270.001
16330.001
17530.001
18450.000
1900.000
20350.000
21370.000
2240.000
2330.000
24280.000
25310.000
26490.000
27400.000
28430.000
299-0.000
3051-0.000
3115-0.000
3252-0.000
3355-0.000
3412-0.000
3556-0.000
3610-0.000
3739-0.000
3850-0.000
397-0.000
4059-0.000
415-0.000
4260-0.000
4361-0.000
442-0.000
451-0.000
468-0.000
4719-0.000
4848-0.000
4929-0.000
5036-0.000
5141-0.000
5242-0.000
5332-0.000
5462-0.000
5530-0.000
5644-0.000
5721-0.000
5846-0.000
5926-0.000
6025-0.000
6147-0.000
6223-0.000
6322-0.000
6463-0.000